SVERIGES GEOLOGISKA UNDERSÖKNING

SER. c. Avhandlingar och uppsatser. N:o 531

ÅRSBOK 47 (1953) N:o 2

PETROLOGY OF THE MÖLNDAL- STYRSÖ-

VALLDA REGION IN THE VICINITY

OF GOTHENBURG

BY

PER H. LUNDEGÅRDH

WITH ONE PLATE

STOCKHOLM 1953 KU:oiGL. BOKTRYCKERIET. P. A. NORSTEDT & SÖNER

Con tents.

P ag. General stratigraphic and tectonic features ...... 3 Supra-crustal rocks ...... I o Quartzite, sla te gneiss, and associa ted rocks ...... IO 1 :\feta-basites, alk ali ne gneiss, cotnmon gneiss, and associated rocks ...... 16 Porphyrite with p henocrysts of plagioclase ...... 30 lnfra·crust al rocks ...... 32 Davainite, gabbro, dioritc ...... 32 Granites ...... 38 Pegmatite, aplitc, vcined gneisses ...... 47 Basic dike rocks ...... 50 Early diabase ...... 50 Late diabase and dolcrite ...... 51 Geoche mical evidences ...... 54 Summary of gene ti cal conceptions ...... • ...... 56 Literature cited ...... •...... 58

Gener·al Stratigraphic and Tectonic Features.

In a recent paper (P. H. Lundegårdh rgsr), I have described the petrology of the Onsala peninsula south of Gothenburg in Western Sweden. I have pointed out that this peninsula is essentially composed of gneisses and gneissic gra:~ it es . Furthermore. the occurrence of various greenstones, sedimentary rocks, and pegmatites has been reported. Some twenty years ago and earlier, the bed-rock of South-Western Sweden as a whole was distinguished as the >> iron-gneiss>formation on account of a con­ siderable content of magnetite in certain varieties (part of the acid and alkaline gneisses). During its evolution, the >>iron-gneiss>> formation has been subjected to a number of tectonic and metamorphic actions. It is thus not surprising that highly divergent opinions regarding its genesis have appeared from time to time. H. E. Johansson (r924 and rg3r) was inclined to define the »iron-gneiss>> formation as folded and fl attened or rumpied products of one single magmatic differentiation. H e did not, however, give any reliable interpretation of the mechanism of such a differentiation. Consequently, other opinions regarding the genesis of the >> iron-gneiss >> formation grew strong. Contrary to Johans­ son's hypothesis, these opinions also take into consideration the supra-crustal rocks (lavas, tuffs, and clastic sediments) that have been discovered in various parts of the formation. 4 PER H. L"CN DEG2\.RDH.

Nowadays, we presume that a series of exogenic and endogenic processes have given origin to the bed-rock of South-Western Sweden. Tbese processes involve weathering of pre-existent rocks, transport, sedimentation, volcanism, folding, and faulting (including overthrusts of blocks and sheets), as well as infra-crustal alterations owing to heat supply, such as re-crystallizations, migrations of ions, metasomatism, dissolutions, and magmatic activity. A short summary of most of our present petrological knowledge of South-Western Sweden will be found in a modern text-book by N. H. Magnusson (1 949). Considerations regarding the orogenie evolution of this part of Sweden have also been offered by H. G. Backlund (1941 ). W. Larsson has published a de­ t ailed scheme of the bed-rock of Northern Dalsland and South-Western Verm­ land (1947, p . 322) . With certain modifications (see below), the division of the Swedish Archaean as given by Magnusson in his text-book of 1949 and in earlier papers will be used in the following t ext . According to Magnusson's scheme, most rocks of South-Western Sweden, and among these the main rocks of the Gothenburg­ Onsala fold, should be classed as Gothian (Table I). Backlund's Gotho-Kare­ lides comprise Magnusson's older, Gothian, and younger, K arelian eyeles (compare Backlund 1941). In my description of the petrology of the Onsala peninsula (P. H. Lunde­ gårdh 195 1, p . 197) , I mentioned that the supra-crustal rocks might be divided into two series. The early series should earrespond to the oldest gneiss complex of W. Larsson (1947) and has to be defined as a lower Gothian supra-crustal series. The late series should have developed simultaneausly with the Åmål series, which is then an upper Gothian supra-crustal series (Table I). When I mapped the rocks of the Mölndal-Styrsö-Vall da region between Onsala 1 2 and Gothenburg ' , I observed that the distribution of the various supra­ crustal rocks (mainly gneisses and amphibolites) spoke in favour of their division into two series. Further supports were provided by the petrographical charac­ t ers of certain members of the supra-crustal series. I have, however, not yet been able to p r o v e the correctness of such a splitting of the supra-crustal series. In the petrological map (Plate r) , the same colours and signs have therefore been used for volcanics and sediments which have developed and altered similarly, the age problem having been ignored. Basic intrusive rocks - gabbros, primary diorites, diabases and dolerites - will be found now and t hen in the Mölndal-Styrsö-Vall da region. Gabbros are the rarest of these, and secondary diorites (re-crystallized volcanics) have proved to be much more common than primary orres. Both groups of rocks are Gothian, whereas the diabase and dolerite cutting the bed-rock as dikes have intruded later. The big dikes running towards W-WNW seem to be Algonkian, whereas the northern to north-eastern dikes belong to the Karelian era.

1 I n this work, I was now and the n assisted b y Mr L. Bergström and Mr J. Lundqvist. The pe trological and structural maps have been prepared for re productia n by Mrs E lisabeth Björk. 2 Theparish of V. F rölunda W of Mölndal (see Plate r) belongs to the territory of Gothenburg. PETROLOGY OF THE MÖLN DAL-STYRSÖ- VALLDA REGION . 5

With few exceptions, the granites S of Gothenburg should be classed as late Gothian. Older granites may be included in the veined gneiss E of Mölndal church (see Plate r ). This rock has suffered from strong metamorphism, how­ ever. Accordingly, it has as a rule lost its primary characters. Late Karelian

Table 1. The Rocks of the M ölndal-Styrsö·Onsala region in order f'rom youngest to oldest. The corresponding rocks of Northern B ohuslän and Dalsland' will be found to the right.

Möl nd al·Styrsö-Onsala region E ra Northern Bohuslä n and D alsland

Diabase and d olerite (W- W N W )

Late pegmatite Bohus pegmatite Youngest gra nite B oh u s gran i te Diabase (N-NE) Koster diabase a nd h yperite

Middle pegma tite, aplite, secondary microline eyes, veined gneiss Askin1 granite2 Kroppefjell Microcline-grani te' gran i te Frölunda gra nite3 Intermedia te gra nite Plagioclase-gran i te' Åmål Basic gra n i te' granite Quartz-diorite, second ar y Diorite, seconda r y Basic and ultra-basic pluta nie rocks Diorite -z, fVolcani cs (compare below)' Volca nics (campare below) ~ - ~ Acid alkaline gneiss• -~ Felspar-quartzite o. 55 Congla mera te (rare) 55 Pie montite-quartzite ~ ~ \Quartzite, sandstone7 Conglamerate 2 ~ Acid volcani cs ;;; Acid volcanics E j E Intermedia te v olca ni cs ...: I n termedia te volcanics u Basic volca ni cs Basic volcanics

Early pegm atite, aplite, Pegmatite, vei ned gneiss ve i ned gneiss Gneiss-gr anites Gneiss-granites Basic and ultra-b asic Ultra-b asic pluta nie plutonic rocks rocks

~ :~ Slate gneiss -~ .. JCi a y sla te ~ ~ §'~J Basic, intermediate and c (J) "' ~ la ctd volca m cs bD ""a - ..":' 2 Quartzi te ~ § lQuartzi te ~ E B asic volcanics ° Basic and acid volca ni cs u 6

1 Northern Bohuslä n according to A. Gavelin: Yttrande med anledning af H. E. J oha nssons föredrag om svenska k varts- och fältspatförekomster, G. F. F., Bd 36, Stockholm 1914, and B. Asklund: K osteröarna, e tt n yckelområde för västra Sveriges geologi, S. G. U., Ser. C, N :o 517, Stockholm 1950. Northern Dalsland according to W. Larsson: Några resultat av berggrunds­ geologiska studier inom Dalformatione ns norra gränsområde, G. F. F., Bd 69, Stockholm 1947, a nd P etrological m ap of the Vår vik region in Northern D alsla nd, printed in 1949, S. G. U., Ser. Aa, N :o r 87. (The description of the map has not been published yet.) F or the most part secondarily magm atic. 3 Most freque ntly second ar y rocks in situ . ' Include porphyrite with p henocr ysts of plagioclase. ' The mutual age rela tions of the rocks included are uncertain. 6 In Gothenburg this rock is sometimes intimately associated with arkose. 7 Long a nd thi n columns of secondary hornblende h ave locally developed here. 6 PER H . Ll.: N DEGÅRDH.

F ig. r. Stra tifi ca tion a nd schis tos ity of the rocks of the Mölndal·Styrsö-Vallda region. Scale I : I 50 000.

För publicering godkä nd i Rikets allmänna k arl\·er k el en 2 r /3 1953. granite has been recently discovered at Näset in the parish of V. F rölunda (see Plate r ). It seems likely that we have to distinguish between three generations of pegmatite. The first of these has developed in early Gothian time, the seeond is older than the Karelian diabase (see above) and should be classed as fin al PETROLOGY OF THE MÖLNDAL-STYRSÖ-V ALL DA REGION. 7

.. ...__..,25-30° MOLNDAL

c

Fig. z. Lineation of t he rocks of the Mölndal·Styrsö·Vallda region. Scale r: r so ooo. För p ubli cering godkä nd i R ikets allmänna kar t verk de n ZI/3 I953·

Gothian, whereas the third one is late K arelian. The veined gneisses are as­ sociated with the first and seeond generations of pegmatite. Most frequently the rocks of the Mölndal-Styrsö-Valida region display schistosities of varying strength, both plane (Fig. r) and linear (Fig. z) ones. The former have become intense along the borders of those blocks and sheets of the bed-rock that have moved in relation to each other as a result of t he 8 PER H. LUNDEG.-\RDH.

final stress of the late Gothian tectonization (compare Flate r). This period of overthrusts has most probably been rather immediately followed by tensional activity: faults and glidings resulting in a migmatization (the seeond Gothian veined-gneiss epoch in Table r; compare P. H. Lundegårdh r951, p. r88 ff.) . Further movements have certainly occurred in K arelian time, simultaneously with the deformation of the supra-crustal Dal series. In Northernmost Dals­ land, W. Larsson (1947) has observed veined gneiss and associated pegmatite that are younger than the Dal series. According to W. Larsson, these late Karelian migmatiiation products belong to the same phase of rock evolution as the Bohus granite and pegmatite, though the paJingenie magma of the latter has developed in deeper parts of the migmatization zone (W. Larsson r947, Fig. r ). The dip of the faults S of Gothenburg varies between 25 and 85° vVNW 1 - WSW (ej. Fig. r and Flate r ). The lineation (Fig. 2) thus frequently runs paraHel to the directions of the movements of the sheets and blocks. Moreover, it coincides with the general direction of the greatest overthrust in Halland and Vestergötland. The border of this steep dislocation can be followed from the Kungsbacka fjord (P. H. Lundegårdh 1951, Flate r) to Lake Mjörn and Lake Venern and has thus a north-eastern orientation. In my Onsala paper, I have shown that both the plane and linear schistosities of the Gothenburg­ Onsala fold developed during the overthrust epoch. There, I have also mentioned that parallels to this interesting coincidence between overthrusts and lineation, are known from the Caledonian mountain range. The late Gothian tectonization probably came into action simultaneously with the formation of the uppermost layers of the late Gothian supra-crustal series (Åmål series etc.; see Table r). In the region where we now find Gothen­ burg, a sinking syncline of both early and late Gothian sediments and volcanics was then transformed into a Iong and rather narrow fold, the axis of which has rotated so that it dips rather rapidly towards NW- WNW. With many excepbons, inter alia due to local distorbons within the fold, the direction of this axis coincides with the lineabon (compare mutually Figs. r -2 and Flate r as well as Figs. r7- r8 and Flate I in P. H. Lundegårdh 1951). As a consequence of the rotation mentioned, the Onsala peninsula represents a seebon through the bottom of the synclinal fold (see P. H. Lundegårdh 1951, Flate r), whereas the upper parts of the fo ld can be studied N of Gothenburg. Its total extension in Bohuslän is unknown, however. E of it, we meet with other folds of Gothian age. In Karelian time, wide parts of South-Western Sweden (Bohuslän and Dalsland especially) were once again subjected to oro­ genie processes.2 As mentioned above, both diabase, granite, and pegmatite have then intruded into the Gothenburg-Onsala fold. The earliest Gothian rocks will be found in the outermost parts of the fold. In the centre, we meet with the bulk of the late Gothian granites, which have developed by transformation of pre-existent rocks (mainly sediments

1 Compare also Plate r and F igs. r7- r8 in P. H . Lundegårdh rgs r. ' See \V. Larsson r947. PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION . 9 and volcanics) during the folding of the syncline and most of which have become tectonized themselves by the stress of the overthrust epoch. As early as 1929, N. H. Magnusson pointed out that the various rocks of the synclinal basin of Gillberga in South-Western Vermland received their present structures simultaneously. This means that both early Gothian gneisses, younger supra-crustal rocks belonging to the Åmål series, and late Gothian plutonites (Åmål-Kroppefj ell granites and associated basites) have become schistase at the same time. The Gillberga basin forms the north-easternmost part of a folded stroke of late Gothian rocks, which can be followed southwards as far as to the Onsala peninsula. The Gothenburg-Onsala fold should thus have developed simul­ taneausly with the Gillberga basin. In Karelian time, however, part of the Gothian formation became covered with sediments and volcanics that have also been folded and faulted. Most of these supra-crustal deposits belong to the middle Karelian Dal series, which was in late middle and late Karelian time subj ected to foldings, disjointings, and overthrusts. The deformation men­ tioned has been effected by an eastern- western stress. W. Larsson (1953) has found that in the Vårvik region (Northernmost Dalsland) the Karelian fold axis shows a distinct , slightly dipping, northern- southern orientation (per­ pendicular to the stress), whereas the fold axis of the Gothian rocks W- SW of the Vårvik region dips slowly westwards. Rather slow, easterly dips are displayed by the fold axes of the rocks E-ESE of the Vårvik region. In Northernmost Dalsland, the stroke of Karelian tectonization has thus to be considered as rather narrow (see W. Larsson 1949). In the Gillberga basin, the lineatian points towards NE or SW, while in the Gothenburg-Onsala fold the direction is most frequently NW. Although both regions are situated in the same northern- southern stroke as the Dal series, they have thus remained rigid during the K arelian orogenesis and their Gothian structures have been preserved.

In 1950, B. Asklund published a summary of his petrological investigations on the Koster and Väder (Weather) isles in Northern Bohuslän. H e there distinguishes between the following generations of rocks: r - an early supra­ crustal, leptitic series, 2 - an early infra-crustal series composed of green­ stones, various non-porphyritic gneiss-granites, and a final gneiss-granite with red or grey, coarse microcline eyes, 3 - diaschishe dike rocks of a lamprö• phyric type, 4 - a late infra-crustal series comprising gabbroic rocks and granite, 5 - Koster diabase and dolerite, 6 - Bohus granite and pegmatite, 7- younger eastern- western dolerite, and 8 - Permian rhomben porphyry. All the rocks of groups 2 and 3 are supposed to be magmatic and to derive from the same parental magma. A coneardant formation of sheets of heterogeneous basites and gneiss-granites compose the older part of the early infra-crustal series. Asklund finds that this contorrnous band-and-lens architecture speaks in favour of the conception of a magmatic differentiation in situ as the origin of these rocks. \ Ne have obviously here to deal with a non-adualistic reasoning IO PER H. L Ul'\DEG.:tRDH. of the same kind as was formerly applied by H. E. Johansson (see 1931 and p. 3) . The data given indicate, indeed, that both amphibolitized basic vol­ canies and supra-crustal rocks granitized in situ have been included in the early infra-crustal series, though of course secondary magmatic activity has largely contributed to the architecture now visible (see for instance Figs. 8- 9 in Asklund r950). When we campare the early infra-crustal series of the Koster isles with the late Gothian infra-crustal series of the Gothenburg-Onsala fold, we find a striking parallelism, whereas the late infra-crustal series of the Koster and Väder isles displays quite another character. Asklund's reasoning indicates, however, a pre-Gothian, viz. Svionian, age of the early Koster series.

Supra-Crustal Rocks. Quartzite, Slate Gneiss, and Associated Rocks.

In the south-eastern part of the Onsala peninsula, at the harbour of Gott­ skär and northwards from there (see P. H. Lundegårdh 195 r, Plate r), the dominant rock is a mixed femic gneiss. When the gneiss was deep-folded in late Gothian time, the bulk of it undenvent plastical deformation and meta­ somatism. At the harbour, we can study one of the results of these alterations: minor penetrative masses of secondary diorite and quartz-diorite. Here, we can also see how the most acid and most basic rock components of the mixed gneiss have escaped the alterations mentioned and display alternating thin layers of white quartzite and green black, skarny davainite (secondary horn­ blendite) . Though they are not very extensive owing to dislocations during the late Gothian orogenesis (Fig. 3), the layers border sharply upon each other and are on the whole remarkably well-preserved. When studying the rocks of Gottskär, we also study the outermost part of the ancient Gothian syncline mentioned in the introduction. The quartzite and davainite layers as well as the surrannding gneiss should thus belong to the early Gothian supra­ crustal series. During my mapping of the isles outside the St yrsö archipelago (Plate r ), I discovered a great number of quartzite remnants in the peculiar plagioclase­ porphyrite of Vinga and Koholmen (p. 30). The xenoliths areangular and border sharply upon the surrounding rock, from which they have become separated by thin sheils of hornblende (Fig. 16, p. 31). As Vinga and Koholmen are also situated in the outermost part of the Gothenburg-Onsala fold, the quartzite remnants may be supposed to derive from an early Gothian series of supra­ crustal rocks. Far to the north, in the archipelago of Northern Bohuslän, we find a similar rock on the isle of Slängerumpan. B. Asklund (1947, pp. 40- 4r) reports that the oldest bed-rock of Slängerumpan is an amphibolite dipping 30-50° SW. This bed is covered by a quartzite layer the thickness of which may sometimes a mount to roo m. Asklund presumes that a continuous layer of quartzite has PETROLOGY OF THE MÖLN DAL-STYRSÖ-VALLDA REGIOK. II

I'ig. 3· P art of disjointed and dislocated sheet of interstratifi ed qua rtzite a nd davaini te in mobili zed basic gneiss. At t he h arbour of Gotts kär, the Onsala penins ula. Photo b v P . H. Lu n­ cl egårdh 1951. once existed near the western border of the Bohus granite. The quartzites of Vinga, Koholmen and Gottskär have probably developed simultaneausly with this layer. ~ot far E of the quartzites, we find a stroke of clayish sandy slates , which I have followed from the Solberga-Marstrand region N of Gothenburg to the Styrsö archipelago. At Tjuvkil and Rörtången, two villages between Marstrand and Solberga, the slate is well-preserved and only weakly folded (Fig. 4). The metamorphism grows stronger and stronger the more closely we approach the Styrsö archipelago, however. (That is to say: the more closely we approach the bottom of the fold considered.) In the St yrsö archipelago, the acid components of the siates have most frequently been assembled to white or grey-white coneardant veins and wind­ ing layers. These are as a rule medium-grained, or even coarse, contrary to the rest of the rock, which has remained fin e-grained. Numerous veins, glands, !enses, and intrusions of red-grey-white pegmatite probably containing juvenile silicate compounds are also found all over the slate gneiss region outside Gothenburg, as evident from Fig. 5 and Plate I (the red winding signs that have been plotted on the yellow colour of the supra-cmstal gneiss). We have seen that the veined slate gneiss is the principal rock of the Styrsö r z PER H. L U NDEGÅRDH.

Fig. 4· Clayish sandy slate. R örtån gen, p arish of Solberga, N of Got henburg. Photo by P . H . Lundegå rdh r g so.

archipelago. Plate r shows that its extension southwards cannot be calculated, as the bed-rock has there been covered by the sea. Veined derivatives of slate gneiss are included in the outer parts of the Onsala peninsula, however (P. H. Lundegårdh r gsr, Plate r), though their areal distribution is here much lower than outside Gothenburg. On the whole, the eastern part of the Gothenburg­ Onsala fold is very poor in derivatives of clayish slates, as compared with its western part. In fact, the charader of the sedimentation seems to have been different in the western and eastern parts of the ancient syncline. E of the Styrsö region, though still in the western part of the fold, several remnants of slate gneiss without pegmatite veins have been observed in the Frölunda granite (the principal rock of the Näset-Fiskebäck-Frölunda penin­ sula; see Plate r ). Moreover, the Frölunda granite itself has developed by al­ teration in situ of strata of clayish sandy siates (see p. 40). From the Onsala peninsula, only few finds of slate gneiss without pegmatite veins have been reported (see P. H. Lundegårdh rgsr, pp. r68- 6g). The naked fresh bed-rock of the isles SW of Gothenburg with their low hills, smooth shelves and shallow valleys is well apt to petrological investigations. From my pocket-books, I shall now quote a few descriptions of typical veined slate gneisses. In the central parts of Vrångö, S of Styrsö, the rock in question displays PETROLOGY OF THE MÖL N DAL-STYRSÖ-VALLDA REGION . 13

Fig. s. Veined and folded slate gneiss. Isle of Vrångö, parish of Styrsö. Photo by P. H . Lunde­ gårdh rgso. folded and frequently distorted layers of variable acidity. Thus, white or grey-white, usually granitic (sometimes aplitic) layers essentially camposed of rather coarse individuals of quartz and feldspar, alternate with grey, inter­ mediate, or black-grey, rnatic and basic layers, which are as a rule fine-grained. The acid layers are more common than the intermediate and basic ones. During the migmatization, the acid layers have been plastic and in part they have dissolved, whereas the intermediate and basic layers, though flexible, have remairred more intact. All over the area considered, the rock is filled with the various kinds of pegmatite injections and segregations described above and displayed in Fig. 5· The coneardant beds of meta-basites now and then observed in the veined gneiss have been rather rigid. Xenoliths of amphibolite derived from disjointed beds1 are thus as common as preserved layers, and the boudinage shown in Fig. 6 may serve as an illustration of the tectonization of these ancient vol­ camcs. In the southern part of Köpstadsö, NNE of Styrsö, the slate gneiss strata frequently wind to and fro. The rock is dark grey to grey and basic to intermedi­ a te or sometimes even acid. The felsic layers have most often been mobilized or at least plastically deformed. When they re-crystallized, they grew coarser and frequently developed an aplitic or granitic character. The mafic layers

1 Compare P. H. Lundegårdh rgsr, Fig. 14· PER H. LUNDEGÅRDH.

Fig. 6. Disjointed sheet of amphibolite in acid to intermediate gneiss. Isle of Vrångö, parish of Styrsö. Photo by P. H. Lundegårdh 1950. have remained fine-grained and contain most of the biotite of the gneiss, whereas the museavite is Concentrated in the felsic layers. The slate gnei::s is filled with glands, lenses, bands, veins, and real intrusions of red-grey­ white pegmatite sometimes accompanied by aplite. In part distorted layers of amphibolite and amphibolitic gneiss have been observed now and then. Under the microscope, the dark fine-grained layers of the veined slate gneiss display a granoblastic re-crystallization texture including the following minerals: quartz, oligoclase, biotite, smaller amounts of museavite and ac­ cessoric quantities of zircon, apatite, titanite, and magnetite. The pale layers are most frequently medium-grained and are apt to show granitic structure. They are composed of quartz, microcline (frequently perthitic), oligoclase, muscovite, and biotite. The former kind of mica is more common than the latter. Accessoric minerals are titanite and apatite. Sometimes, garnet has also been found, whereas, owing to the sandy charader of the mother sediment, no minerals extraordinarily high in aluminium, such as cordierite, sillimanite etc., have developed. The pegmatite and aplite of the veined gneiss will be described later (p. 47 ff.). I have already mentioned that amphibolitic rocks are rather frequent in the slate gneiss region. Before the Gothian orogenesis, these meta-basites displayed coneardant beds of basaltic lava, tuff and tuffite. At present, their mineral composition and textural-structural development do not refl ect many of their primary features, however. They are mainly camposed of acid plagio­ clase (most frequently oligoclase) and common hornblende, further of some quartz, biotite, apatite, and, though not always, magnetite (often titani- PETROLOGY OF T H E :vrÖLNDAL-STYRSÖ-VALLDA REGION . I5

Fig. 7. Banded gnei ss: interstra tified a mphibolitic gneiss and more acid gneiss with porph yro­ blasts of microcline. NW of Långåker , parish of K ållered. P h oto by J. L undqvist I95I.

ferous), epidote, and titanite. Except for the larger individ u als sometimes preserved (the phenocrysts of the ancient lava), the plagioclase has crystal­ lized as granoblastic grains. All the quartz has developed similarly. In a thin section of amphibolitic gneiss from the southernmost part of Donsö (see Plate r ), the hornblende shows the following pleochroic colours: a - colourless or very pale greenish-yellowish, f3 - olive-green, and y - blue-green. z Vy amounts t o about 95° and c:y to 14- 15°. Other rocks associated with the veined slate gneiss are sedimentary leptites (fine-grained, granoblastic) and mixed or banded gneisses (Fig. 7). Inter­ mediate to basic gneisses of tuffitic origin are sometimes also found. A dark grey leptitic slate derivative from Stora Rävholmen near SW of Styrsö has been investigated microscopically. Owing to alterations, its primary stratifica­ tion cannot always be recognized. The main minerals are quartz, oligoclase, and biotite. Garnet is quite common, too, and should be classed as an inferior constituent. Accessoric minerals are magnetite, apatite, muscovite, zircon, and allanite. Of the minerals m entioned, only garnet and biotite to some extent follow the layering and thus make this structure visible. Small aggregates of quartz, most frequently lens-shaped, have been met r 6 PER H. LUNDEGÅRDI-I.

F ig. 8. Sheet of schistase congtornerate in plagioclase·granite. Isle of Rivö, parish of S t yrsö. P hoto b y P . H. Lundegårdh rgsr. with now and then all over the slate gneiss region. It seems likely that this q uartz has been once liberated during the successive alteration of the clayish sandy sediments and that it has moved to positions of minimal pressure during the folding. Strokes and masses of acid, intermediate and basic granitic rocks are quite common in the slate gneiss region, especially on Rivö and Styrsö (see Plate r ). Sometimes, these rocks have developed by granitizations in situ, sometimes by intrusions of palingenic (secondary) granitic magma. Both processes seem to have occurred in late Gothian time. The western parts of Stora Känsö and Vargö (see Plate r) essentially consist of granitized slate gneiss. Numerous schlieren (remnants of folded slate layers) very rich in mica here reveal the sedimentary origin of the granite. On Rivö W of Fiskebäck (see Plate r), a grey plagioclase-granite has intruded in to the slate gneiss. At the strait between Rivö and Asperö, a sheet of schistase conglamerate (Fig. 8) has been observed in this granite. The conglamerate displays pebbles of re-crystallized and rather coarse quartzite in a matrix of mafic gneiss. The age of the conglamerate will be discussed on p. 24.

Meta-Basites, Alkaline Gneiss, Common Gneisses, and Associated Rocks. E and SE of the Styrsö archipelago, the Gothenburg-Onsala fold contains a lternating layers of met a-basites and gneisses derived from volcanics and sediments. These have become in part distorted, in part destroyed by tectoniza­ tions, granitizations, dissolutions, and magmatic intrusions. Owing to the high degree of met amorphism of the supra-crustal rocks mentioned, their age and PETROLOGY OF THE MÖLNDAL- STYRSÖ-VALLDA REGIOK . rJ

Fig. g. Amphibolitic gneiss with a la yer of intermediate gneiss that h as allered to Askim granite. NW of Lå ngåker, parish of Kå llered. Photo b y J. Lundqvist 1951. primary charader may seem difficult to discern. The positions of the various strokes of ro.:;ks in the fold serve to indicate the general stratigraphy of the supra-crustal series, however (compare Table r) . Thus, the innermost rocks are youngest, viz. late Gothian. These expose the picture of disjointed steep beds of meta-basites covering a kernel s tro k e of alkalin e gneiss (see P late r). Among the former, preserved volcanics have been recently discovered, at first on the Onsala peninsula (P. H. Lundegårdh rgsr, pp. r63- 65). In rare cases, even the alkaline gneiss is associated with rocks whose primary structures have been preserved (in Gothenburg arkose; on the Onsala peninsula ag­ glomerate, see P. H. Lundegårdh rgsr, p. r67) . The remaining gneisses of the rnainland part of the Mölndal-Styrsö-Vall da region are most probably in part early, in part late Gothian. As was mentioned in the introduction, the division of the supra-crustal rocks in two series is not definitive, however. To a considerable extent, the gneisses of the Mölndal-Styrsö-Vallda region have been granitized in situ. Thus, the alkaline gneiss has been now and then transformed into microcline-granite, while the intermediate and basic gneisses have given origin to intermediate and basic granites (compare the Frölunda granite mentioned above). On the other hand, the greater part of the inter­ mediate, porphyritic Askim granite, and the plagioclasic, frequently rather basic granites, have developed by crystallization of secondary (palingenic) magmas, viz. deep-folded, liquefied and mobilized rocks. The same mode of development holds for part of the intermediate, non-porphyritic granites. On the whole, we can say that both modes of granitization now discussed have frequently intermixed. 2-53083-l. S. G. U., Ser. C, N:o SJI. Lttndegårdh. 18 FER H. LUNDEGÅRDH.

As regards granitizations in situ, the following general statements have to be considered. H. G. Backlund (1936, 1941, 1943) has repeatedly pointed out that, during the evolution of the cm st, there has been n o l a c k o f t i m e f o r l a r g e - s c a l e g r a n i t i z a t i o n s, further that g r a n i t i z a­ tions can proce ed und e r th e condition of eonstant v o l u m e, owing to the migrations in opposite directions and resulting re­ placements effected by the ions involved in the alteration process. Moreover, L Th. Rosenqvist (1949, 1952) has stated that alterations of this kind have most probably been mainly effected by i n t e r g r a n u l a r m i g r a t i o n s o f d i s s o l v e d i o n s. The metamorphic basic volcanics, or meta-basites, of the Mölndal-Styrsö­ Vallda region can be divided into three groups, two of which will be described here. These groups are amphibolite (meta-basite with more than so % horn­ blende) and amphibolitic gneiss (meta-basite with less than so % hornblende though still rich in this mineral). The latter may pass into mafic gneiss. lt has proved impossible to distinguish between the amphibolite and the amphibolitic gneiss during the map-work (compare Plate I), as both rocks have the same appearance. The third group, the secondary quartz-diorite and diorite, is described in the next chapter. Originally, the beds of basic volcanics to a very great extent alternated with layers of more acid composition, as a mle tuffites and epiclastic sediments (for instance quartzite, see p. 10). This interstratification has been preserved in a number of cases, as is seen from Figs. 7 and 9· Most frequently, it has been distorted, however. During the Gothian orogenesis, the intermediate and acid strata became plastic and movable. In part, they were even dissolved and could then behave like magmas. Typical results of these processes are displayed in Figs. ro- r r. Fig. r z shows an interesting though very local p henomen on: tensional joints filled with mobilized gneiss. \iVhen interstratifications of inhomogeneous supra-cmstal rocks have been preserved, we speak of banded gneisses, when distortians have occurred, we use the term mixed gneisses . The amphibolitic gneiss and amphibolite are grey-black, black or green­ black, fine-grained rocks. Most frequently, they have become influenced by the regional tectonization, though they are usually less schistase than other supra-cm stal rocks of the Gothenburg-Onsala fold. As a rule, their present t exture has developed secondarily. In certain cases, however, plagioclase pheno­ crysts have been preserved from the original volcanics. The main constituents are common hornblende (secondary after pyroxene etc.) and plagioclase (oligoclase or andesine). Variable quantities of quartz, biotite, and epidote are also found. In the amphibolitic gneiss, the former belong to the main minerals. Chlorite is sometimes an important constituent, too. Most of the epidote and chlorite seems to have developed during the final Gothian migmatization (the seeond veined gneiss epoch in Table r ; ej. p. 38), whereas the principal alteration of the basic v okanies is a result of the deep-folding of the Gothenburg-Onsala syncline. In order to make the mineralogical picture of the meta-basites complete, PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 19

Fig. ro. Basic to amphibolitic gneiss penetrated by mobilized intermediatc gneiss. stensholmen (isle t NW of Särö), parish of Släp. Photo by P. H. Lundegårdh rgsr. we have to mention the eonstant presence of a small amount of apatite. Mag­ netite (frequently titaniferous) is another minor constituent that ought to be reported. In rare cases, this mineral is lacking, however. The hornblende has crystallized as columns, prisms, or irregular individuals. The latter are frequently camposed of aggregates of small grains. Certain individuals have sometimes grown coarser than the rest of the rock, thus giving it a porphyritic appearance. Repeated development of long and thin hornblende columns (sometimes even needles) is characteristic of strongly lineated areas. Most columns are, of course, then approximately paraHel to the lineation. The plagioclase (except the sparse phenocrysts) and the quartz are as a rule granoblastic. In the meta-basites, some few strakes and masses of preserved volcanics have been observed (campare P. H. Lundegårdh rgsr, pp. r63-65). Im­ 1 mediately E of the highway I / 2 km NNW of Kållered church (Plate r) , for instance, I have encountered a sill of grey-black to black dolerite1 camposed of labradorite, common green hornblende (secondary), clinopyroxene, biotite (secondary), magnetite, and red garnet (secondary). Furthermore, a minor content of apatite (small rods) should be reported. The labradorite forms primary laths (ophitic texture) and flocks of grano­ blastic grains. The margins of the former have become corroded during the

1 Spectral a nalysis in Table I II, discussion of the data obtained on p. 54· 20 PER H. LUNDEGÅRDH.

Fig. r r. I ntrusions of mobilized acid gneiss in amphibolite. MaJeviksholmen (isle NW of Särö), p aris h of Släp. Photo by P . H. Lundegårdh rgsr. alteration of the rock. A few primary individuals show zonal extinction. The clinopyroxene, probably an augite (y 1\ c = 40°), has become strongly im­ pregnated with dispersed magnetite. The secondary hornblende has either developed as separate granoblastic grains or penetrates the pyroxene along margins and fissures. The biotite has most frequently assembled to aggregates of minor individuals. The garnet is penetrative. In the mountains 2 km ENE of Vallda church (inferior locality) and 2 km NE of V. Frölunda church (superior locality), a fine-grained, grey-green-black, amphibolitic gneiss with oval pebbles of pale grey quartzite and grey-white sandstone has been discovered (Plate r and Fig. IJ). The sandstorre pebbles are mainly composed of quartz and oligoclase. The length of the pebbles is 1 1 1 / 2- r / 2 cm. The gneiss should be classed as a conglamerate with tu ffitic matrix. It is associated with a series of folded and in part disjointed beds of late Gothian basic tuff and lava. The tuffitic matrix1 of the conglamerate NE of V. Frölunda church is essen­ tially compose:l. of oligoclase and hornblende (first-rate constituents), biotite and quartz (second-rate minerals), epidote and titanite (inferior minerals). The minor constituents are apatite, allanite, magnetite (probably titaniferous), and pyrite. The t exture is secondary: xenomorphic and in part granoblastic. The hornblende is a uralite with pale pleochroism in blue-green (!') and grass- ' In part, the matrix mav b e a ltered lava (see p. 32 ). PETROLOGY 01' THE ~fOLNDAL-STYRSÖ - VALLDA REGIOK. 2I

Fig. rz. Amphibolite pene tra ted by mobilized acid gneiss along tensional joints. N orth-western H osholme n (isle SW of Särö), parish of Släp. Photo by P . H . Lundegårdh 1951. green ((3). 2Vy amounts to goo or samewhat more. The ailanite is, in thin see­ tians, brown-yeilow and has become surrounded by hornaxial epidote. Owing to alteration, it is most frequently isotropic·. As mentioned above, the pebbles of the conglamerate have become oval. This deformation has been effected by the stress of the late Gothian tectoniza­ tion. Simultaneously, the tuffite has grown schistose (most frequently parallel to the stratification). As already mentioned, the pebbles consist of either quartz alone or qnartz and oligoclase. Minor minerals are epidote, hornblende, allanite (isotropic), titanite, biotite, oxide and sniphide ore. In part, these emanate from the tnffite. In part of the conglomerate, the qnartzite pebbles have become endosed in sheils of hornblende. The thickness of the sheils is variable, and some pebbles have been completely amphibolitized. The sig­ nificance of this phenomenon wiil be disenssed on p . 32. The pebbles seem to derive their origin from a sedimentary rock camposed of alternating layers of quartzite and sandstone. Snch a rock has also been found r km SW of Askim church and on the southernmost part of Sillfarsholmen, an isle between Köpstadsö and St. Förö (in the western vicinity of 'St. Förö' in Plate r). N of the Fiskebäck-V. Frölunda area (outside the northern border of Plate r), it is rather common (see H. E. Johansson rg3r, p. 28) . When in­ vestigated microscopically, the sedimentary rock of Sillfarsholmen displays a strong epidotization of the qnartz-oligoclase (sandstone) layers, whereas the 22 PER H. LCN DEGJ-'tRDH.

F ig. 13. Basic tuffite with pebbles of quartzite and sandstone. Natural size. NE of V. Frölunda church, at the northern border of Plate L Photo b y C. Larsson 1952.

quartzite layers have remairred intact. Furthermore, the rock is traversed by long, thin columns of common, secondary hornblende (Fig. r4) . These have developed before the epidotization (compare p. 23) and have then become in part altered to penninite. The primary stratification is well-preserved. It has been made visible inter alia by the distinct layers of pure, granoblastic quartz, inter alia by several grains of magnetite (probably titaniferous and frequently associated with titanite). The epidote of the ancient sandstorre layers, too, has b een distributed parallel to the stratification. This mineral has crystallized as an immense number of minor individuals, many of which are idiomorphic. The remairring oligoclase, on the other hand, is xenomorphic and even apt to show granoblastic texture. The quartzite-sandstone rocks SW of Askim church and outside the northern border of Plate r have altered similarly. Thus, for instance, they contain the same penetrative columns of secondary hornblende as the sillfarsholmen sediment (compare H. E. Johansson I 9JI, p. 28). Both on Sillfarsholmen and SW of Askim church, the formation of hornblende has started from faces that are at the same time planes of stratification and schistosity. The number of horn­ blende columns is still highest along these faces. Furthermore, on Sillfars­ holmen, the formation of hornblende is dependent on the presence of basic volcanics in the close vicinity of the quartzit ~ -sandstone rock. From the vol- PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 23

Fig. I4· Quartzite-sandstone rock with p orphyroblasts of h ornb lencle. Scale 4 : 5· S illfarsholmen (isle t b etween Styrsö a nd Fiskebäck ), parish of Styrsö. Photo by C. Larsson r952. canics, magnesium and iron have been introduced along the mutual planes of stratification and schistosity. Metasoroatic reactions have then taken place. On a small scale, these imply a basification of part of the sediment (the horn­ blende columns) and an acidification of the nearest vakanies (an increase of the quartz and felspar content). Structurally, the hornblende-bearing quartzite-sandstone rock is reminiscent of the Garbenschiefer of the Caledonian mountain-range. Similar rocks have also been found in the late Gothian supra-cmstal rocks of the Gillberga basin in south-Western Vermland (N. H. Magnusson rgzg, p. I3}- Furthermore, 'vV. Larsson (1949) has observed a few strakes of rocks with thin columns of secondary hornblende SW of the Gillberga basin. Larsson reports that even late Gothian granites, such as the Åmål granite, have there been subjected to secondary formation of columnar hornblende. If we consicter the process mentioned to have occurred at the same time all through the Gothian bed­ rock of South-Western Sweden, we have to place it in the seeond Gothian period of migmatization (see Table I). During the Karelian orogenesis, the quartzite-sandstone rock does not seem to have been subjected to any more considerable heating. (Campare the problematic and in any case restricted low-temperature metasornatism effected by the late Karelian magmatic ac­ tivity in the southern vicinity of Gothenburg, for instance at Näset, p. 47.) The penninization of the columnar hornblende and the simultaneous epidotiza­ tion of part of the quartz-oligoclase layers may, however, have taken place either in late Karelian time or when the influence of the late Gothian mig­ matizing agents faded (c/. p. 38). On the other hand, we have in Northern Dalsland and South-Westernmost Vermland no visible migmatization in latest Gothian time, whereas considerable alterations were eaused by the Karelian orogenesis. As touched upon in the introduction, the Dal series of sediments 24 PER H. LUNDEGÅRDH.

and volcanics grew thick, sank, and became t ectonized together with the underlying Gothian bed-rock. According to W . Larsson (r947), the Bohus granite and pegmatite should be classed as the final results of a late Karelian migmatization that has given origin to veined gneiss and earlier pegmatite at higher levels of the crust , for instance in Northernmost Dalsland. W. Lars­ son (personal communication) is therefore inclined to interpret the hornblende porphyroblasts SW of the Gillberga basin as products of this migmatization. As the Gillberga basin has been far less influenced by the Karelian orogenesis, the porphyroblasts here met with may be older (late Gothian). The quartzite-sandstone rock seems to belong to the lower group of lat e Gothian sediments and volcanics. Outside it, we have, in the fold, metamorphic lavas and tuffs most probably corresponding to the basal part of the Åmål series (see Table r ). Before the folding, these have covered the early Gothian Marstrand-Styrsö slate. The general conformity between the late Gothian vol­ canies mentioned and the early Gothian slate, bear evidence of a rather quiet evolution of the Marstrand-Gothenburg-Onsala region in early and middle Gothian time (compare Fig. 4). As the conglamerate has derived all its pebbles from the quartzite-sandstone rock, its interformative character can be considered as indisputable. In Table r, it has thus been placed closely above the rock mentioned. The position of the conglamerate on Rivö (p. r6) is more uncertain, however. This conglamerate also displays a higher frequency of quartzite pebbles than the rnainland one, and its position in the fold indicates an early Gothian age. The quartzite-sandstone rock and the tuffitic conglamerat e seem to be the only wholly definable sediments of the inner part of the Gothenburg-Onsala fold. There are to be found, however, much more extensive strokes of rocks, the sedimentary origin of which cannot possibly be doubted when their earn­ positions are studied. These rocks now display layered granoblastic gneisses of variable acidity. The stratification is the single primary structure observed. It has, however, most frequently become intensified during the late Gothian tectonization, especially in and along the faults of the region investigated. The dominant gneiss is a rather acid, felsic, red-grey to pale red variety - the renowned alkaline gneiss of the Gothenburg region. Together with the late Gothian volcanics (basic tuffs, tuffites, and lavas), this rock constitutes the innermost layers of the Gothenburg-Onsala fold and continnes in the centre of the fold N of Gothenburg, as evident from H. E . J ohansson's petro­ logical map of the Kungelv-Gothenburg region (H. E. Johansson rg3r). The Gothenburg stroke of alkaline gneiss has been faulted paraHel to the stratifica­ tion and schistosity so that, N of Mölndal, its eastern part disappears from the present surface of the bed-rock (compare mutually Plate r and the petrological map in H. E. Johansson rg3r). P araHel strokes have been found E of Gothen­ burg, as exemplified by H. E. J ohansson's m ap. S-SSW of the Kållered area (Plate r), the Gothenburg stroke of alkaline gneiss has lost its continuity owing to distortians effected by the evolution of the late Gothian granites. In part, the alkaline gneiss has there even altered into microcline-granite. PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 25

The main minerals of the alkaline gneiss are microcline, 1 quartz, and oligo­ clase or oligoclase-albite. As a rule, the former are first-rate constituents. NW of Kållered church and from there towards Gothenburg, the leading mineral is always microcline (see the geometric analyses given below). Though it is far less common than the plagioclase, biotite may also be counted a main mineral. Among minor constituents, I shall at first mention those that are nearly always present, viz. magnetite, titanite, zircon, and apatite. Mag­ netite is at times ~·ery common and in certain cases displays coarse individuals. Owing to their frequent content of magnetite, the alkaline and acid gneisses of South-Western Sweden were formerly distinguished as 'iron gneiss'. In this name, all rocks associated with the alkaline and acid gneisses were also incorporated and thus we got the 'iron-gneiss' formation (compare the in­ troduction). iitanite is sometimes common, too (campare the geometric analyses given below). In and N N'vV of the K ållered area, hornblende rich in iron appears among the minor minerals and may even replace biotite and magnetite (see below). Garnet and fluorite have also been observed here. WSW- S of Kållered church, epidote enters the alkaline gneiss and grad­ ually replaces the hornblende. The mafic minerals have most frequently been arranged so as to make the stratification of the rock visible even in detail. A remarkable exception from this rule is given by the hornblende, however. I have already mentioned that the content of hornblende may grow high in the north-easternmost part of the Mölndal-Styrsö-Valida region. Indeed, I have been able to distinguish in Flate I a stroke of alkaline hornblende gneiss. Even the amphibole of this rock is alkaline. 2 It displays secondary prisms and short columns that have crystallized as solitary individuals of variable orientations. Simultaneously, the original minerals have re-crystallized and grown larger. The ancient sedi­ ment has thus obtained a rather granitic habitus. In thin sections, however, the primary stratification may still be traced, though it has frequently been destroyed during the growth of the secondary crystals. The typical hornblende gneiss of the stroke distinguished in Flate I is a pink or reddish white, medium- to fine-grained rock camposed as follows: microcline > quartz > oligoclase-albite > black hornblende > garnet, titanite, and zircon. (The three last-mentioned are minor minerals.) A geometric analysis of the alkaline hornblende gneiss 3 km NW of Kållered church has given the following data (volume-% ): :Vli crocline ...... 46. o Q uartz ...... 29 . 5 Oligoclase-alb ite ...... 18. 4 Alkaline h ornblende ...... 6.o Garne t, titanite, zircon ...... 0.2

I OO. I

~ felsic minerals = 93.8 % ~ mafic n1ineraJs = 6. z o/0

1 Perthitic interlamin a tion of acid plagioclase is uncommon. 2 I n the Gothenburg region, H. E . J ohansson (1931, p. 34) has observed alkaline clino­ pyroxene, too. z6 PER H. L UNDEGÅRDH.

For comparison, a sample of the common alkaline gneiss 2 km W of Mölndal .church has also been subjected to geometric analysis (volume-%):

Microcline ...... 44· 1 Quartz ...... 29.5 Oli goclase ...... r 8.8 Biotite ...... • ...... 3· 8 Titanite...... r. s Common hornblende ...... r. o Jl!agnetite ...... o.S Garnet ...... o. 3 Zircon, ilmenite e tc...... o . 2 100. 0 ~ felsic minerals= 92.4 % ~ mafic minerals= 7.6 %

The relative and absolute contents of the three felsic minerals are thus about the same in both rocks. All the minerals of the rocks are fresh, and no indications of tectonization have been observed. In view of their exposed positions in the Gothen burg-Onsala f old (see P late r ), i t is therefore evident that the rocks considered cannot possibly have obtained their present appearance before the final Gothian migmatization. Indeed, I am inclined to interpret it as a product of this process. The hornblende of the quartzite-sandstone rock earlier described should thus have developed simultaneausly with the amphibole of the alkaline gneiss, and both should be products of metasornatism (campare below). The hornblende of the common alkaline gneiss shows normal pleochroic colours (y - deep blue-green, f3 - deep olive-green), whereas the amphibole of the alkaline hornblende gneiss has quite another pleochroism: y (sometimes a) - deep blue, f3 - deep blue-violet, a (sometimes y) - pale green-yellow, or grey-green, or medium green-yellow. The extinction angle is very small. In most individuals, y(\c falls between o and 5°. The birefringence is here mode1ate, viz. o. o 1 z a o. o 14. A minority of individuals (with inverse pleo­ chroism as indicated by the parentheses above) show a(\ c = o-5° and very weak birefringence. It is quite obvious that we have here to do with two members of the glaucophane-riebeckite series. Most of the crystals examined have to be classed as crossile (intermediate stage between glaucophane and riebeckite), whereas the weakly birefringent individuals show optical properties that are characteristic of riebeckite. The composition of crossite approximately earresponds to the following form ula:

In riebeckite, the iron content has increased:

(OH) 2 Na2Fe3++ Fe2 ++ + Sis022 Both sorts of amphiboles are thus rich in sodium and iron. The former meta! has been supplied by metasomatism, the latter derives from the mag­ netite that is now lacking. The magnesium of the crossite is likely to have come from pre-existent biotite. Two chemical analyses of red, granitic, fine-grained alkaline gneiss from Gothenburg have been published by P. J. Holmquist (rgo6, pp. z68-6g): PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGIO!\. 27

1 Fig. 15 . Intense selective weathering of alkaline gneiss. z / 2 km W of Mölndal church. Photo b y G. Lundqvist 195 r.

I. S of Johannis' II. At Vega church street Si02 •••••• .. • •••• ••• • •• • • • • • • • 75·34 75.69 TiO, ...... o. l 5 0.16 Al2 0 3 •• ••• ••• ••••• • • •• • • •••••• I2 . 5 I II.64 Fe2 0 3 •••••••• • •••.• • •••• • ••• •• o.6z I. J O FeO ...... r.sz 0.84 MnO ...... o. 23 0.2 I i\1g0 . .... • ...... 0.2 0 o. z o CaO ...... 0 . 4 0 o. 75 Na2 0 ...... · · ·· ·· 2 .00 2.42 K 2 0 ...... •...... 6. 55 6. 16 P, O ...... · . · · 0. 0 4 H,O ...... 0.36 0.66 gg.SS roo. o7 In the southern, south-western and western slopes of many hills and moun­ tains, the hornblende-bearing alkaline gneiss of the Gothenburg-Mölndal region has been and is still being subjected to a rather intense mcchanical weathering of the same kind that has given origin to the name of the Finnish 1 Rapakivi granite (Plate I and Fig. rs; see also the Swedish geological 2 map-sheet 'Göteborg' ). There seems to exist a positive earrelation between a lkalinity and disposition of weathering, as already suggested by H . E. Johansson (1931, pp. 34- 35). As is displayed by Fig. rs, the weathering of the alkaline gneiss works 1 Rapa kivi = rotten stone. 1 Sveriges geol. undersöka., Ser. Aa, N:o 173, Stockholm 1931. 28 PER H. L UND EG.Å. RDH. selectively. Part of the gneiss has thus escaped disintegration and remains as. ovoids or sheets in the mass of sliding gravels. The remaining gneisses of the rnainland part of the Mölndal-Styrsö-Vall da region expose to view a spectrum of basic and intermediate and even acid, re-crystallized and often granitic varieties. The general mode of occurrence (interstratifications etc.) and principal kinds of alteration of these common gneisses have already been described. The basic gneiss essentially consists of plagioclase, quartz, common hornblende, and biotite, while the intermediate and acid varieties are most frequently characterized by an increasing content of microcline. Simultaneously, the hornblende has been seen to disappear gradually or suddenly. In most cases, the intermediate and acid vari eties are also richer in quartz and poorer in plagioclase than the basic ones. A t ypical exponent of magnetite-bearing acid gneiss will be found in Table II. The plagio­ clase is an oligoclase or, in the basic varieties, sometimes an oligoclase­ andesine. Titanite, magnetite (often titaniferous), and apatite are the minor minerals most frequently met with, though zircon and pseudo-allanite are also characteristic accessories, at least in the intermediate and acid varieties. The occurrence of muscovite seerus to be more accidental. (The rnainland slate derivatives are then not considered.) Garnet is quite rare, whereas epidote is the leading secondary mineral of t he common gneisses. In certain strokes, it belongs to the main minerals. The same must also be said of the late Gothian granites described in the next chapter. In the above text , we have seen that epidote is a first -rate constituent of the quartzite-sandstone rock, too, and that it is common even in the basic volcanics. Epidote is a typical low-temperature mineral. My investigation of the Askim granite at Gottskär (P. H . Lundegårdh rgsr, p. r Sr ff.) has shown that it is secondary and post-tectonic. Most frequently, it appears as aggregates of flocks of idiomorphic grains. At Gottskär, the epidote seerus to have devel­ oped simultaneously with the microcline porphyroblasts of the granite. As I am inclined to interpret the latter as products of the final Gothian migma­ tization, I should alsolike to place most epidotization at the end of the Gothian era, in spite of the finds of Karelian granite and pegmatite S of Gothenburg. The presence of garnet and, in the alkaline gneisses, hornblende, too (see above), as a rule excludes the occurrence of epidote. As the former are high­ t emperature minerals, this might indicate a contemporaneous development of all these minerals. Indeed, considering the rnainland part of the Mölndal• Styrsö-Vallda region, the north-eastern corner is w i t h o u t d o u b t c h a r­ a c t e r i z e d b y a n o t h e r m i n e r a l f a c i e s than the rest of the region. Compare, for instance, the garnet-bearing supra-cmstal dolerite NNvV of Kållered church (p. rg) with its chemical equivalents in the Särö-Vallda region. In the latter, the primary clinopyroxene has a rule been totally altered and instead of garnet we find epidote. Compare also once again the hornblende­ garnet-bearing alkaline gneiss NW of IZållered church with the epidote-bearing alkaline gneiss SSW of Kållered. It is significant that in the gneiss at K ållered PET ROL OGY OF THE :\1ÖLNDAL-STYRSÖ-VALLDA REGI O N. 29

Table II. Chemical analyses of rocks from the Mölndal-Styrsö-Onsala region. Analyst: A. Aaremäe.

. B asaltic Pl . 1 U 1. Gra nite, As!cin1 D 1 · l Kind of r ock G ne1ss, l t ff 'i agtoc ase -l ra t te- l t! l . l o en te grey, acid lu ' agt: -porph yrite -porphyrite greyb, ra ler l granite (dike) l g on1era 1c astc ' 1 At the At the South- E:E ir. s kmE ofl+-s km ES; I highway 75 highway eastern 2 km vV N vV fmSl" lth e railway of t he rail- L ocalit y 2 . 5 km o ap . f l . N W of 1 SSW of part of of Meryt h h s tatlon o lway station. Mer y t' Vinga c urc Billdal of Billdal l Gottskärl l l l

72- 34 48· 97 59 · 47 46. ;6 63-39 66.8 8 49· l O o. 39 I. 2 5 1.77 0 .47 o. 72 o . ;8 z.g6 I 2. 94 I 5· 6 I 14. l I 7 . l 6 I 6. t 4 I 5 · 3 5 I7. l 8 0 .47 I.67 I. 99 3· 28 I. 77 0-99 I. ;6 2.48 8. t 6 7. 20 6. 59 3.68 3. 0 2 g .63 o. o6 O. l 7 0 .1 2 0. l 4 o.og o. o6 0. 1 2 ...... o.64 7. s 2 2 . I I I 6.94 2.05 r. 26 4.68 I. I 6 g. 54 4-70 I 4-94 4-76 3· s+ 7-3 7 0. 07 O. OJ o. os 0. 03 o. o; o. o; 0. 04 5-74 I. 8 8 3· 24 0. 47 2.40 3· JO I. 3 7 Na,O ... 2.34 2. 3 8 3.02 o. 55 3-4 3 3-26 3-6 5 P, o • ...... 0 .1 I 0. I 8 o.;z o.o8 0 . l 6 o. l 5 0.48 H,O > I 10° . .. 0 .97 2 . J I 1.49 2.8 I I. 39 I. I O I. 27 s ...... o. o; o . 28 o. og o .o8 o. t 8 o. 23 0 , 1 I F ...... 0. l 9 o. og o. t 6 0 . 0 l 0. 0 2 0 . 2 0 O. OJ o. o8 0 . I 8 o. o8 O. II o. '4 o. ' 4 O. l ) ---~------l S um l JOO. OJ 100.2 2 100 . [ 2 100. 2 2 I oo. 3 7 100. I l 99-90

Oxygen replaced by sulphur a nd fluorine ...... O . I O o. J 5 0. J O 0. 04 o. o8 0. l 7 0. 04

99· 93 100. 07 100. 0 2 100. l 8 100.29 99-9+ gg.86

l See P . H . Lundegårdh 1951, P late I.

both hornblende and epidote belong to the ordinary constituents, whereas garnet is already lacking. Pressure conditions being favourable, garnet and 1 epidote can join, however. The gneiss 2 / 2 km SSW of Gottskär (see Table II), for instance, contains both the low-temperature mineral epidote and the high­ pressure mineral garnet, though the latter only occurs as sparse grains in the mafic lay ers of the rock. The slate gneiss stroke of the Styrsö archipelago is also free from epidote, whereas garnet is sometimes met with. Furthermore, the slate gneiss seerus to have become migmatized in latest Gothian time, viz. at the same time as the gneissic rocks of Eastern Mölndal. The slate gneiss stroke and the rocks of the north-eastern corner of the region investigated can therefore be said to have been simultaneausly subjected to the same kind and degree of met a­ morphism. The rapid variations of metamorphism within the Gothenburg-Onsala fold have already been touched upon in my Onsala paper (P. H. Lundegårdh 30 PER H. LUNDEGÅRDH. rgsr). I there considered t hem to be indicative of the existence of t wo supra­ cm stal series. According to this opinion, we should have to deal with preserved associations of secondary minerals developed by altering agents of different ages. The above interpretations have, however, shown that, in the Gothenburg­ Onsala fold, two mineral facies have been able to develop simultaneausly in different parts of the same rock. Furthermore, the mode of occurrence of the acid gneiss rich in alkaline hornblende bears evidence of sudden changes of the degree of one and the same alteration process. Examples of similar rapid intensifications of regional metamorphism are displayed by the charnockites of Central and Southern Halland (P. Quensel r gsr). As these also belong to the Gothian evolution of rocks, they may, indeed, have attained their present mineral facies simultaneausly with the development of alkaline hornblende in the Gothenburg-Mölndal region (Quensel rgsr, pp. 315 and 318).

Porphyrite with Phenocrysts of Plagioclase. The naked isles of Vinga, Koholmen (immediately S of Vinga), and F järskär (z km ESE of Vinga; see Plate r) expose a supra-crustal greenstone of peculiar chemical composition (Table II; see also Table III and p. ss). The weathered surface of the rock has a pale reddish-grey hue (on F järskär grey-red) that made me expect to meet with a granite when I first landed at Vinga. The yellowish-white felspar eyes immediately reminded me of phenocrysts, how­ ever. A great number of weil-preserved xenoliths (Fig. r6) soon confirmed the correctness of my primary impression. Indeed, the porphyritic greenstone seems to have wholly escaped secondary alterations, and no signs of tecto­ nization except jointing have been observed. In fresh specimens, the Vinga greenstone displays a fine-grained greenish black-grey matrix essentially camposed of common hornblende with sparse remnants of clinopyroxene and oligoclase (at times tabular) with alteration products, sericite especially (campare the high potassium value in Table II) but sometimes calcite and clinozoisite, too. Further, the matrix, which is xeno­ morphic, contains much titanomagnetite, quartz, and microcline. Minor minerals are penninite, biotite, and apatite (numerous rods). The felspar phenocrysts are in part t abular and also camposed of oligoclase. 1 F resh pheno­ crysts have a greyish-greenish-white hue. On the whole, oligoclase is the leading mineral of the rock. The hornblende of the plagioclase-porphyrite is optically negative (zV ]arge)' and pleochroic in various shades of green. It has developed as aggregates of small individuals showing irregular shape and variable orientation. It is as­ sociated with titanomagnetite, at times with biotite and penninite, too.

1 When my thin sections of rocks from the Mölndal-Styrsö-Vallda region were examined, a basic porphyritic xenolith filled one section completely and therefore happenecl to be inter­ mixed with the ordinary plagioclase-porph yrite. The p lagioclase of the xenolith proved to be andesine to labradorite, and thus the plagioclase-porphyrite was at first prese nled to Swedish petrologists as p orphyrite with phenocrysts of labradorile (Geol. föreningens i Stockholm för• hand!., Bel 74, p. 531, Stockholm 1953). PETROLOGY OF THE M ÖLN DAL-STYRSÖ- VALLDA REGION . 3I

Fig. r 6. Plagioclase-porphyrite with xenolith of qua rtzite endosed in thin shell of hornblende. lsle of Vinga, p arish of Styrsö. Photo by P. H . Lundegårdh 1952.

In the clinopyroxene, a diallage, y f, c amounts to about 43°. As mentioned,. most clinopyroxene occurs as remnants in the hornblende. Single minute crystals have, however, also been observed in the oligoclase. Obviously, these have been saved from alteration thanks to the surrounding felspar. The quartz and microcline are the latest primary minerals of the plagioclase­ porphyrite. Their mother liquor, a t ypical residual solution, has rnaved in the intergranular film, and their replacement of pre-existent minerals has st arted from there. Among the phenomena developed by this process, I shall mention myrmekitic intergrowths. Simultaneously, the water and potassium of the rest solution have effected strong autometamorphism and autometa­ somatism. The uralitization of the pyroxene and the sericitization of the plagio­ clase bear witness of this alteration. In connection with the sericitization, the plagioclase has become impregnat ed with minute hematite grains. On V inga and Koholmen, the deuteric sericitization is moderate and the impregnation weak or sometimes even lacking, whereas on Fjärskär, which seems to lie nearer the border of the plagioclase-porphyrite, the oligoclase individuals of the matrix have been completely changed into a sericite mass crowded with hematite grains. Neither the quartz nor the microcline has escaped the hematite im­ pregnation, and most of the rock thus displays a beautifully red colour. Si­ multaneously, part of the uralite and all of the biotite have been chloritized. 32 PER H . L UND E GÅ RDH.

Only the phenocrysts have remairred comparatively intact, inasmuch as their kernels are free from hematite and their sericitization is incomplete. The borders of the plagioclase-porphyrite are covered by the sea. As al­ ready totlChed upon, the rock is, however, very rich in well-preserved angular xenoliths. Pieces of early Gothian quartzite (Fig. r6) are most common, but a great number of the xenoliths display various other rocks such as red felsic gneiss, sandy slate belonging to the Marstrand-Styrsö stroke, met a-basites (compare the fo ot-note on p. 30), and pegmatite derived from some kind of veined gneiss. F urthermore, single remnants of a red helsinkitic rock have been encountered. On the other hand, xenoliths of l a t e G o t h i a n g r a n­ i t e s h a v e n e v e r b e e n o b s e r v e d, although Frölunda grani te (also non-deformed) occurs as n ear as on In-V inga, two small islands lying r km NNE of Vinga (see Pla te r ). I therefore feellike equalizing the plagioclase­ porphyrite with the latest Gothian volcanics, 1 in spite of the fact that no indications of secondary metamorphism have been traced. The xenolithic peg­ matite should then be at least early Gothian. (Xenoliths of Svionian = Pre­ Gothian rocks m a y of course occur in the plagioclase-porphyrite.) Most of the quartzite xenoliths and part of the others have been endosed in thin shells of black hornblende during the solidification of the plagioclase­ porphyrite (Fig. r6) . We have earlier seen (p. zr) , that similar hornblende shells have been found around many of the oval quartzite pebbles of the late Gothian conglomerate. Furthermore, we have observed that some of these pebbles have been complet ely amphibolized. We do not know, whether this alteration is primary or secondary. If it is primary, we have, however, to interpret the amphibolitic matrix of the conglamerate as a basic lava.

Infra-Crustal Rocks. Davainite, Gabbro, Diorite.

In magmatic evolutions, ultra-basic rocks as a rule develop first. It is thus quite natural, that even in the Gothenburg-Onsala fold the earhest products of magmatic differentiation are ultra-basites. These are rare rocks, however. In the Mölndal-Styrsö-Vallda region, we only know three small deposits of ultra-basic magmatic differentiation products, while on the Onsala peninsula five minor masses have been discovered (P. H. Lu.ndegårdh rgsr, p. IJO). A sample from one of latter has been analysed chemically (Table II). In Table I, I have tried to distinguish between two generations of magmatic ultra-basites. The older of these should correspond to the soapstones of Dals­ land (see W. Larsson 1947) and Brattön ENE of Marstrand (near the locality shown in Fig. 4). The younger ultra-basites have intruded into the late Gothian supra-crustal series, probably at the beginning of the late Gothian folding. Both generations of ultra-basite are frequently associated with normal gabbroic rocks.

1 In the Åmål series (Table I ), we .h a,·e also porphyrites. PETROLOGY OF THE MÖLN DAL- STYRSÖ-YALLDA REGIOK. 33

Fig. 17. Amphibolized olivine-gabbro with bands of davainite {meta- peridoti te) . Stensholmen (islet NW of Särö), parish of Släp. Photo by P . H. Lundegårdh 1951.

The ultra-basic magmatic differentiation products of the Mölndal-Styrsö• Vallda region are situated on s tensholmen 2 km NW of the centre of Särö, further 2 km SSE and 2 km S of Vallda church (Plate r). The s tensholmen ultra-basite occurs as paraHel sheets in a metamorphic olivine-gabbro passing into diorite or dioritic amphibolite. The orientation of the sheets is in general NIS0 E, 65°ESE. They usually appear as bands in the outcrops examirred (Fig 17). The ultra-basic sheets of s tensholmen display a black green, felt-like matrix of variable grain. During the late Gothian tectonization, they have become ratber scbistose. They are camposed as follows: hornblende > penninite > magnetite ~ biotite > pyrite """ calcite ~ apatite. They should be classed as davainitic (campare bel o w). The hornblende has developed as larger and smaller, xenomorphic, or some­ times bypidiomorphic (prismatic), individuals which have become intermixecl with penninite all through the rock. In this mass, the remairring minerals have been scattered in a rather poikilitic manner, though they are in part idio­ morphic (the pyrite especially). Much of the calcite has crystallized along fis­ sures, however. In thin sections, two kinds of hornblende have been observed, viz. a common one showing rather strong pleochroism and an actinolitic one that has very pale pleochroic colours. This indicates tbat the original rock has been a lherzolite (a periclotite very rich in clinopyroxene). 3-5-108 .1 -1. S . G. U., Ser. C, N:o 5JI . Ltmdegårdh. 34 P ER H. LUNDEGÅRDH.

The stensholmen davainite has been subjected to spectral analysis (Table III). The data obtained are discussed on p. 55· The davainitic sheets seem to have developed by rhythmic crystallization differentiation of a gabbroic ( = basaltic) magma, though the possibility of metamorphic banding cannot be wholly rejected (campare P. H. Lundegårdh 1946, pp. 73-75) . The interstratified olivine-gabbro has altered to hornblende­ gabbro, and the diorite also consists of plagioclase and secondary hornblende. \Vhen grading or changing in to amphibolite ( = the margin of the basic in­ trusion), the diorite is n ev er banded. The davainite and amphibolitic-dioritic rocks have in part been brecciated by mobilized gneiss (Fig. 18) . The ultra-basic intrusion 2 km SSE of Valida church displays a sheet-like, steep, homogeneous mass of black-green, medium-grained to coarse davainite with a marked t endency to porphyritic texture (hypidiomorphic hornblende porphyroblasts). During the late Gothian tectonization, a great nu m ber of fissures paraHel to the schistosity of the adj acent gneiss and granite have developed. The hornblende grains around the porphyroblasts have then been frequently crushed. Except hornblende, which is the predominant mineral, the davainite SSE of Valida consists of moderate quantities of biotite and smaller amounts of titanite, epidote, apatite, magnetite, and pyrite. Furthermore, the occurrence of sparse individuals of primary plagioclase should be reported. These are almost completely altered t o saussurite, however. Accidental grains of secondary, rather fresh and acid plagioclase have also been observed. The hornblende is a common secondary one with pale green pleochroism. y/\c amounts to about ! 2° and 2Vy to 95- 100°. The ultra-basite 2 km S of Valida church has developed quite similarly though the intrusive body is here more rounded than SSE of Valida and the visible marginal zone rather dioritic or amphibolitic. Mobilized acid gneiss has also repeatedly penetrated and brecciated this davainite. Among the metamorphic basic volcanics of the Mölndal-Vallda region, davainitic rocks have been öbserved in a few outcrops. The areal distribution of these wholly secondary ultra-basites is very restricted. Gabbro is as rare a member of the Mölndal-Styrsö-Vallda rock family as davainite. The metamorphic olivine-gabbro on s tensholmen NW of Särö has already been mentioned. In the rnainland part of Plate I , gabbro will only 1 be found in the mountains 3 / 2 a 4 km S of Mölndal church and 2 km ENE - ~ E of V. Frölunda church. The gabbro S of Mölndal is a strongly met amorphic, greyish or greenish black, fine- to medium-grained rock which passes into quartz-diorit e. The only primary constituent of greater importance is a corroded t abular plagioclase, most of which has altered t o sericite. The rest of the rock essentially consists of common hornblende that seems to be secondary after clinopyroxene, of granoblastic oligoclase, and of biotite. Granoblastic and in part poikilitic quartz is frequently also a rather important mineral. With increasing content of quartz, the gabbro passes into quartz-diorite. Minor minerals are apatite and PETROLOGY OF T H E MÖ L ~DAL-STYRSÖ-VAL LDA REGIO~. 35

Fig. r8. Amphibolite and davainite brecciated b y mobilized intermediate gneiss. Stensholmen, parish of Släp. P hoto b y P . H . Lundegårdh rgsr. titanite. The latter is especially common in the dioritic varieties. P yrite is the only accidental constituent met with. The gabbro 2 km ENE-NE of V. Frölunda church has developed quite otherwise. A single look at this lens of grey(greenish)-black, rather fine-grained biotite-norite endosed in an amphibolitic uralitization shell (Plate r ) will be enough t o distinguish it from the greenstorres above described. Genetically, the biotite-norite belongs to the Slottskogen1 stroke of hyperitic dolerite and norite and amphibolitic derivatives described by H . E. Johansson (1931, pp . 32-33). The Slot tskogen basite has in part been disjointed during the late Gothian t ectonization, as evidenced by t he lens 2 km ENE-NE of V. F rö• lunda. The mineral composition of the biotite-norite of this lens is as follows: plagioclase > biotite ~ augitic diaHage 'Y quartz > uralite > magnetite ~ hypersthene 'Y apatite (minor mineral). A spectral analysis is given in Table III. The data obtained are discussed on p. 56. The plagioclase is sometimes tabular but m ore often xenomorphic. The largest individuals show zonal extinction (margin = 30 % anorthite, kernel = 50 % anorthite). On the whole, the plagioclase is an andesine (35 % anorthite or a little more). Now and then, the plagioclase individuals have in part altered to sericite.

1 ;\Iunicipal p ark in South-Western Gothenburg. PER H. L UNDEGARDH.

The biotite forms !arge primary packs of sheets showing beautiful pleo­ chroism in red-brown. These packs are frequently associated with magnetite. The augitic diaHage is weil-preserved though in many cases margirrally ural­ itized. z Vy amounts to about 65° and yl\c to about 45°. As a rule, it constitutes fairly small individuals and at times aggregates of rounded grains, wherea~ the hypersthene has developed as somewhat larger and less xenomorphic individuals. The hypersthene has also remairred rather intact. The quartz is late and penetrative. M?rmekitic intergrowths have been met with in single cases. The uralite is a common hornblende pleochroic in various shades of green. The Slottskogen stroke of noritic basites obviously continnes far south­ wards, though i t hasthere become disjointed and as a rule completel? uralitized, too. At present, it isthus characterized by dioritic and amphibolitic greenstones that most frequently cannot be distinguished from the metamorphic basic 1 volcanics of the region. 3 / 2 km SE of Billdal, bowever (Plate r ), an eastern­ western, dioritic and in part meta-noritic, xenolitb measuring 8oo m in length has been found in intrusive Askim-granite that locally passes into plagioclase­ granite. \V. Larsson, who has investigated the late Gotl1ian diorite of Northem­ most Dalsland (Table r), has told me that this rock has the same general appearance as the greenstorre SE of BilldaL The mode of occurrence of both rocks is also similar (compare mutually Plate r and W. Larsson 1949) . 1 Most part of the xenolith 3 / 2 km SE of Billdal displays an amphibolitic quartz-bearing diorite, but now and then and especially in the westernmost part of the xenolith, the meta-noritic character is quite obvious. A thin section from here shows a greenstorre consisting of plagioclase = uralite > magnetite ~ quartz = biotite > epidote ~ apatite (minor mineral) ~ pyrite (accidental mineral). The plagioclase contains about 50 % anorthite (andesine-labrador:ite) and has in part crystallized as hypidiomorphic, lath-shaped individuals. Frequently, it has undergone partial sericitization. Flocks of minute grains of epidote, in part idiomorphic, are also common. The quartz of the rocks is late, interstitial and penetrative. Some plagioclase individuals grow acid when bordering upon quartz. The uralite shows zVy about 95 ° and variable yl\c. Its pleochroism is green and most frequently, especially in larger individuals, margirrally strong and centrally weak. Oriented microlites of ilmenite occur now and then. Further­ more, single remnants of augite must be reported. Most of the uralite is evidently secondary after this mineral, though pseudomorphs after hypersthene have been sometimes observed, too. As already mentioned (p. r8), the leading greenstorres of the Mölndal-Styrsö• Vallda region are secondary, amphibolitic and dioritic rocks (see also Plate r). Most of the latter have been shown to derive their origin from the same kind of basites as the former and have therefore to be classed as mere alteration products of basic volcanics (compare P. H. Lundegårdh 1951, p. 174). These dioritic rocks have developed during the late Gothian folding, simultaneous l~- PETROLOGY OF THE :IIÖLNDAL-STYRSÖ-VALLDA REGIOX. 37 with the late Gothian granites. Thus, even regarding their age, they do not differ considerably from the majority of those rather few dioritic rocks that can be proved to have an infra-crustal magmatic origin, for inst ance the rock 1 3 / 2 km SE of Billdal (see above) . The only distinction is the slightly higher age of the latter (see Table r) . These st at ements evidence that transitions between dioritic and amphi­ bolitic rocks are very common (compare also Plate r and P. H . Lundegårdh rgsr: Plate r). They also tell us that no mineralogical differences do as a rule exist between the t wo groups of basites. The diorite and quartz-diorite thus essentially consist of plagioclase, common hornblende, biotite, and quartz (main mineral only in the latter kind of rock). Furthermore, epidote is an important constituent that at times must be classed as a main mineral. It is frequently accompanied by considerable quautities of chlorite. Minor con­ 1 stituents are apatite, titanite (which is lacking in many samples of diorite ) , and magnetite (most often titaniferous) . Now and then, pyrite has also been observed. The dioritic basites display grey, grey-green, black-green, or greenish-grey­ black, fine- to medium-grained rocks. The plagioclase, an andesine (in the quartz-diorite sometimes oligoclase-andesine), is in part tabular, whereas the remaining minerals (except the epidote) are xenomorphic. In samples of primary diorite, the plagioclase may show zonal extinction. Most frequently, the mineral has in part altered t o sericite, and in many cases it has also become filled with minute, idiomorphic or hypidiomorphic crystals of epidote. The hornblende is pleochroic in various shades of green. zVy as a rule lies between 95 and I 00°, y/\c between I5 and 20°. Poikilitic inclusions of quartz have been found in many samples. The hornblende has most frequently de­ veloped as aggregates of grains including most part of the other mafic minerals, though single grains and larger, penetrative individuals also occur. The quartz is late, penetrative, and therefore often poikilitic. Yl ost of the epidote and chlorite of the dioritic rocks seem to have developed during the final stage of the late Gothian migmatization and, in any case, contemporaneously with the epidote of the late Gothian granites (see p. 38). Partial or total re-crystallization of some of the dioritic (and amphibolitic) rocks are also likely to have occurred during this epoch, though of course somewhat earlier than the crystallization of the low-temperature minerals just mentioned (compare the alteration of the Sillfarsholmen quartzite-sandstone rock, p. zr). Indeed, we have in certain kinds of diorite (for instance on Sill­ farsholmen) large schiliering intergrowths of hornblende which have developed later than the remaining amphibole and, at least a part of the plagioclase of the rocks in question. The occurrence of such greenstones twice altered seems to be as accidental as the occurrence of hornblende porphyroblasts in the quartzite-sandstone rock and alkaline gneiss (see the foregoing chapter).

1 Titanium is as a rule concentrated in late prod ucts of differentiations of basic magmas (see P . H . Lundegårdh 1950). PER H. LUXDEGARDH.

Granites. The Gothian granites of the Swedish West Coast can be divided into two groups, viz. early Gothian and late Gothian granites. The former seem to be rare in the Mölndal-Styrsö-Vallda region. Indeed, they have only been traced in the veined gneiss strake E of the Gothenburg-Mölndal-Kungsbacka fault. It is, however, difficult to distinguish them there from supra-crustal gneisses, and they have thus not been marked on Plate r. In the region considered, the late Gothian granites constitute the dominant group of acid infra-crustal rocks. They have been painted brown in Plate r. They have to be classified as follows : r. Black-grey to dark grey, basic granite most frequently rich in hornblende. This granite frequently passes into quartz-diorite. 2. Red-grey to grey, intermediate granite (Frölunda granite and plagioclase­ granite). J. Red-grey to red, rather acid granite (microcline-granite). 4· Grey-red to dark red-grey granite with coarse eyes of microcline (Askim granite) . The late Gothian granites are most frequently rather gneissic. This structure is generally due to tectonization and appears as plane and linear schistosities (Figs. r -2). The resultant rock is thus a typica1 gneiss-granite, or schistase granite. As I mentioned in the introduction, the strength of the tectonization has been highly variable. In Plate r , strakes of granite (and other rocks) showing marked schistosity have been striated. Sometimes, the paraHel structure of the late Gothian granites has another origin. In those cases when the granites have developed in situ by alteration of pre-exist en t rocks (granitization in sit,u), gneissose remnants of the altered rocks are common. A granitized mica schist, for instance, can thus be recognized as paraHel bands, schlieren etc. in the secondary granite, which has then to be classed as granite-gneiss, or gneissose granite. The granite-gneisses can of course also have been subjected to tectonization. Rocks that are at the same time gneissose and schistase can therefore sometimes be encountered. The schistosity may even cut the gneissosity. Moreover,. exaroples are known where dikes of gneiss, mobilized along disjointing planes of schistosity during the final Gothian migmatization, cut schlieren and sheets of remnant gneiss and meta-basite in secondary granites (especially in various parts of the coast region between Fiskebäck and Kullavik; see Fig. 23). The basic granile is concentrated in the south-eastern part of the Mölndal• Styrsö-Vallda region (Plate r). It displays a black-grey to dark grey, fine- to medium-grained, locally non-deformed but far more often gneissic rock (quartz­ dioritic gneiss-granite). Main minerals are plagioclase, quartz, biotite, common hornblende, and, in most samples investigated, epidote, too. As has been mentioned earlier, the epidote has developed secondarily, in most part of the region probably during the final Gothian migmatization (see p. 28) . Around PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 39 the late Karelian rocks in the north, epidotization also seems to have occurred in late Karelian time, however (p. 47). Now and then, this mineral has replaced all the hornblende, for instance at Släp church (the basic granite analysed in Table II). It has cryst allized as small individuals which frequently expose well-defined faces and are then rod-shaped. I t does not only form interstitial individuals but has also crystallized onto the biotite. Further, it appears as flocks of minute idiomorphic and hypidiomorphic grains in the plagioclase (which has simultaneausly lost part of its lime). The latter should be defined as an oligoclase-andesine or basic oligoclase, that has frequently re-crystallized to acid oligoclase, or in rare cases even to oligoclase-albite, during the epidotiza­ tion. Partial sericitization is another common kind of alteration. The larger plagioclase individuals are often roughly t abular (primary hypidiomorphic development). Plagioclase is most frequently the leading mineral of the basic granite, which also repeatedly passes inta intermediate plagioclase-granite. The xenomorphic or hypidiomorphic hornblende shows normal pleochroism in various shades of green. z Vy as a rule falls between go and I00°, yAc between 0 r8 and zo • The minor minerals are the normal orres of basic granites, viz. titanite, titaniferous magnetite (in part probably titanomagnetite), and apatite. In the plagioclase, zoisite is sometimes met with, and the biotite may locally have altered to penninite. Late microcline appears now and then, frequently as porphyroblasts. Single grains of zircon, allanite (in part pseudo-allanite), and pyrite have also been observed in many samples of basic granites. As touched upon above, a sample of rather basic granite from Släp church has been analysed chemically. Epidote belongs to the main minerals of this rock and has there replaced all the hornblende. Furthermore, it has aquired part of its lime from the plagioclase. The rock is now camposed as follows: oligoclase > quartz > biotite > epidote > titanite > apatite > pyrite. Apa­ tite and pyrite are minor minerals, whereas titanite has occupied a samewhat stronger position. Most of the basic granite seems to have developed on crystallization of a secondary magma (see the plagioclase-granite below). The intermediate granite comprises two different varieties, viz. the Frölunda granite, which is rich in microcline, and the plagioclase granite, which is as a rule poor in, or free from, microcline. Furthermore, museavite is rather frequent in the former. The Frölunda granite is Concentrated in and N of the Fiskebäck-Näset-V. Frölunda area, where it is the dominant rock. It displays a grey, or at times red-grey, medium- or fine medium- to fine-grairred rock, which has most frequently escaped tectonization. The Frölunda granite consists of quartz, oligoclase (often showing weak sericitization), microcline (sometimes perthitic), and biotite (second-rate main mineral). Museavite and secondary epidote (see the basic granite) are rather important constituents, too. Titanite, apatite, and sometimes also magnetite should be classified as minor minerals. The Frölunda granite is in most cases quite homogeneous. Now and then, PER H . LUND EG..:\RDH. however, it displays alternating, parallel, straight or slightly winding sheets of variable composition. All gradations exist between hardly recognizable in­ homogeneity and distinct banding, as can be seen both on and near the coast bet ween Fiskebäck and Näset (Plate r ). Among banded varieties there en­ countered, I shall mention a rock camposed of alternating layers of dark grey, rather basic granite and red grey, rather acid granite. ~or e o ve r , the Frölunda granite frequently contains coneardant schlieren, bands, and dike-like sheets of rocks with preserved supra-cmstal charaders. This even holds for homogeneous varieties. I have already mentioned the granite on St. Känsö and Vargö with its schlieren of mica inherited from pre-existent slate gneiss (p . r 6) . I will also eaU attention to the In-iVin ga sles N of Ving a. Fig. 24 displays a detail of the bed-rock on In-Vin ga. The narrow greyish white bands and schlieren are Frölunda granite, whereas the thick grey layers are re-cryst allized, more or less granitic supra-cm stal gneiss. The Frölunda granite is here intermediat e to acid, the gneiss basic to intermediate. On part of In-Ving a, the gneiss vanishes, however, and the Frölunda granite forms rather homogeneous masses. The petrological data now reported show that most of the Frölunda granite has developed by granitization in siitt. Furthermore, the considerable content of museavite and the petrographical character of the supra-cmst al remnants indicate that the pre-existent rock has been early Gothian sl ate gneiss with intercalat ed layers of basic volcanics. Of course it is the most resistant layers of the pre-existent slate gneiss strake and especially then the intercalated beds of meta-basites (amphibolitic and femi c gneisses) that have been preserved as sheets in the secondary granite. Owing to t ectonization during the granitization, these sheets have often been disjointed and now appear as rounded or angular pieces. Fig. 2r shows such an irregular inclusion. However, real eruptive dikes that are older than the post-Gothian diabase and dolerite also occur in the late Gothian granites, as has been already touched upon in my description of the geological map-sheet 'Onsala' (P. H. Lundegårdh r g52, Fig. r o1, p. 29). At Västra Hagen, an important petrological 1 locality situated r / 4 km ~E of the centre of Onsala Sandö (see P. H. Lunde­ gårdh r gsr, Plate r ), such dikes have been shown to derive from gneisses that have become plastically deformed and in part mobilized during the final Gothian migmatization. In the Mölndal-Styrsö-Vall da region, sparse though straight and sometimes rather extensive dikes of mobilized gneiss have also been observed (Fig. 23) . These dikes are as a rule grey, fine-grained, and acid to intermediat e. Further, the existence of dikes of late Gothian paJingenie granites should be reported. Both basic granite, plagioclase-granite, and Askim granite are represented among these dikes. As a rule, they penetrate gneisses and basites bordering upon masses of granite (Figs. rg-2o), but they may at times occur

1 ·w ith the exception of this figure, the paper in Swedish now citecl cl oes not cliffer fro m P. H . Lunclegårclh 195 1. PETROLOGY OF THE :\I ÖL:\TDAL-STYRSÖ-YALLD.-\ REGIO:\T. 41

Fig. rg. Dike of grey plagioclase·granite in q uartz·dioritic a mp hibolitic gneiss. Stora Små· holmen (islet SW of Billd al), parish of Askim. P hoto by P. H . Lundegårdh 1951. even in granitized rocks, such as the Frölunda granite. In the left part of Fig. zr , we observe a dike of grey, rather basic plagioclase-granite in Frölunda granite. As is also shown in Fig. zr, this dike has been cut and dislocat ecl by late Karelian granite. The plagioclase-granite is grey, fine medium- t o medium-grained, and as a rule slightly schistose. Main minerals are oligoclase (sometimes oligoclase­ albite due to secondary development of epidote), quartz, biotite, and in many samples epidote, too (compare the basic granite). The larger plagioclase in­ dividuals are often t abular. Partial sericitization and impregnation with flo cks of small epidote cryst als are common phenomena. Plagioclase is the leading mineral of this rock, whereas hornblende, on the other hand, has inferior im­ portance or is, in normal cases, even lacking. Titanite, muscovite, and apatite have to be classified as minor constituents. P yrite, zircon , and allanite (incl. pseudomorphs) are more accidental. Microcline sometimes attains a strong position. When this mineral begins to form coarse eyes, the rock passes into .-\skim granite (see below). In hand specimens, the two kinds of intermediate granite do not often differ greatly from each other, though the Frölunda granite is on the whole more fine-grained than the plagioclase-granite. They have also been given the same symbol in Plate r. Genetically, the difference between the two rocks is great , however, though they pass into each other in the Askim area. 'vVe have seen that most of the Frölunda granite has developed by granitization in situ. We PER H. L UNDEGA RDH. have also read about dikes of plagioclase-granite, and basic granite, too, in Frölunda granite, in supra-crustal gneisses, and in basites. Pictures of such dikes have also been shown (Figs. rg and zr). Moreover, the plagioclase­ granite and basic granite are usually free from remnants of older rocks that can be interpreted as indicative of granitizations in sit ~t , whereas, on the other hand, genuine eruptive breccias are sometimes met with (see P. H . Lunde­ gårdh I 95I, Fig. 6). Most of the plagioclase-granite and basic granite therefore seem to be magmatic rocks, though their magma was certainly secondary, viz. composed of liquefi ed rocks from the deepest part of the Gothenburg­ Onsala fo ld (see the summary, p. 56) . In certain areas, for inst ance Särö and Vallda Sandö S of Särö, the general character of the supra-cm stal remnants observed is the same as in the Frölunda granite, however, and thus bears strong evidence of a granitization in situ. Transitions between plagioclase-granit e (even basic varieties) and Askim granit e are f o und all over the rnainland part of the Mölndal-Styrsö-Vallda region (see Plate r ). As soon as the former grows rich in coarse microcline eyes, it should be classed as Askim granite. Since many of the microcline eyes are secondary, viz. porphyroblasts, the high frequency of transitions is by no rueans remarkable (see t he Askim granite described below). The red-grey to red, fine- to medium-grained microcline-granite is a gran­ itization product of the same kind as the Frölunda granite, though the paren­ tal rock is, in this case, acid and alkaline gneiss. The int imate relationship between the latter and t he microcline-granite is indicated already by their distribution as shown in Plate I , and the mode of development of the microcline­ granite is elucidated in several outcrops, for instance SSE-S of Särö. The microcline-granite also consists of the same minerals as the acid and alkaline gneiss, viz. microcline, quartz, and oligoclase-albite to oligoclase, further biotite and magnetite. Regarding the latter, which are seeond-rate minerals, it ought to be mentioned that they are usually not simultaneonsly frequent. Considerable amounts of biotite thus depress the content of mag­ netite, or even exclude this mineral, and vice versa. Minor constituents are epidote, muscovite, zircon, titanite, and apatite. The epidote, which is also here secondary (see the basic granite), may sometimes attain a strong position. Remnants of microcline-granite in intermediate andAskim granite have been observed in a few outcrops, for example in the mountains 5 km SE of BilldaL The leading late Gothian granite is the porphyritic .1l skim granite. The matrix of this rock as a rule earresponds to one or other of the granites described above and is consequently quite variable. Most frequently, however, the matrix has the same intermediate composition as a microcline-bearing plagioclase­ granite. The microcline of the Askim granite is thus concentrated to the eyes. There are, indeed, varieties to be found, the matrix of which is free or almost free from potassic felspar. The matrix of the Askim granite is grey-red to grey or dark grey and fine­ t o medium-grairred (occasionally in part coarse), whereas the oval or rectangular microcline eyes are pink or red owing to impregnation with minute hematite PETROLOGY OF THE ~IÖL XDAL-STY R SÖ-VALLDA REGIOX. 43

Fig. 20. Dike of Askim granite in basic gneiss. 2 km NE of V. Frölunda church. Photo by P. H. Lundegårdh r952 .

1 .grains. They usually measure / 2-5 cm in length. Perthitic in tedamination of acid plagioclase is common. Apart from microcline, the main minerals are quartz, oligoclase, biotite, and, in strongly altered varieties, epidote (compare the basic granite). Another important mineral is titanite. The following minor {;Onstituents have been observed: apatite, titaniferous magnetite, and often muscovite, too. Zircon and pseudo-allanite (sometimes allanite) should be dassified as accidental minerals. The tectonization of the Askim granite is highly variable. In the Billdal­ Kullavik area, for instance (Plate r), the rock is often non-deformed (chemical analysis in Table II), but in most cases it has become more or less schistose. The crushed minerals - oligoclase and quartz preferentially, have re-crystal­ lized as granoblastic aggregates. In dislocation zones (striated in P late r) , the schistosity grows very strong. A detailed description of Askim granite thus tectonized will be found in my Onsala paper (P. H. Lundegårdh rgsr, pp. r8o-83). The genesis of the micro­ cline eyes is also discussed there. On the east coast of the Onsala peninsula, most of these existed when the final Gothian tectonization and migmatization started. Only a minority of eyes have proved to be youngerthan the tectoniza­ tion mentioned. The older eyes have often been cracked. The layers or sheets of minerals composing the zones of schistosity also pass around these eyes without interruptions, whereas they have been in part replaced by the pene­ trative microcline forming the post-tectonic porphyroblasts. The latter are 44 PER H. LUNDEGÅRDI-L frequently also idiomorphic = rectangular. The petrological significance of the post-tectonic eyes is great, because they have obviously developed at low temperature. Indeed, any more considerable post-tectonic heating of the schis­ tase Askim granite would have brought about complete re-crystallization. All cracked and smashed minerals would have then disappeared. 500 m SE of Släp church, an instructive outcrop has been found. The pre­ dominant rock is here a plagioclase-granite which eastwards passes into Askim granite. In the border zone, the microcline eyes are concentrated along planes of schistosity that have allowed potassium-bearing solutions to invade the plagioclase-granite. Furthermore, the porphyroblasts are associated with sparse and small, non-deformed dikes of red aplite. Similar dikes have been obsenred at a few other localities, and it may even happen that these late dikes contain microcline eyes. In view of the data now referred to, I am inclined to interpret most of the late microcline eyes of the Gothenburg-Onsala fold as final Gothian, though, in the vicinity of Gothenburg, such porphyroblasts may also have developed in late Karelian time. 1 We have there intrusive rocks produced by the late Karelian migmatization, viz. granite and pegmatite. Furthermore, we have, round the Karelian granite of Näset, areas rich in late microcline porphyro­ blasts, the development of which has eaused the Frölunda granite there occurring to pass into secondary Askim granite (compare Plate r). 2 km NE of V. Frölunda church, just outside the northern border of Plate r and close to an intrusion of late Karelian pegmatite, 2 even arkose and con­ glamerate have been in part transformed into Askim granite by the mere ac­ tivity of granitizing solutions. The conglamerate belongs to the stroke described on p. 20 and has the same general charader, though its matrix is in part less basic, which has thus rendered granitization possible. Accordingly, the con­ glamerat e with amphibolitic matrix has remained intact, whereas the con­ glamerate with matrix corresponding to femic gneiss has been granitized. In this case, however, the alteration observed is most probably late Gothian. Indeed, the great distance between the minor late Karelian intrusions S of Gothenburg and the ancient zone of late Karelian migmatization has certainly prevented the Gothian bed-rock of the Mölndal-Styrsö-Vallda region from any more con­ siderable alteration in Karelian time. Epidotization and chloritization may have occurred (see for instance p. 21), and even microcline low-temperature porphyroblasts may have developed (see above), but complete granitizations of supra-crustal rocks cannot possibly have been effected. The selective granitization just mentioned, implying the preservation of the basic part of a supra-crustal rock, is a normal phenomenon in the Gothen­ burg-Onsala fold. Fig. ro shows a banded gneiss, the acid layers of which have been transformed into Askim granite while the amphibolitic layers only contain sparse microcline porphyroblasts. Investigations on the schistase Askim granite of the Onsala peninsula (see

1 Compare the secondary epidotization earlier disenssed (p. 38). 2 The Högsbo pegmatite, see p. 49· PETROLOGY OF THE MÖLNDAL--STYRSÖ-YALLDA REGIOX. 45 above) and other parts of the Gothenburg-Onsala fold have shown that most of the microcline eyes of this rock are older than the final Gothian tectoniza­ tion. The field work has revealed the absence of remnants indicating granitiza­ tions in sitn in most areas of Askim granite. Developments of secondary Askim granite similar to those just mentioned (granitized matrix and secondary eyes) are thus inferior phenomena. On the contrary, the central and southern main­ land parts of the Mölndal-Styrsö-Vallda region display huge eruptive breccias of supra-crustal rocks in late Gothian granites, among these Askim granite (the Billdal-Kullavik area especially, see Plate r ). Small breccias are also rather frequent (see P. H. Lundegårdh rgsr, Fig. 7). Furthermore eruptive dikes of Askim granite in older rocks are sometimes met with (Fig. zo). The magmatic character of most of the Askim granite is thus quite obvious, though this magma was certainly secondary (see the summary, p. 56). _-\s regards the genesis of the early eyes, I shall mention that small and large inclusions of non-porphyritic acid gneiss (in rare cases even microcline-granite) have sometimes been met with in the magmatic Askim granite, for instance at the sea about z km SSW of Billdal (acid gneiss, see Plate r). If wholly secondary potassic solutions had been at work here, the acid gneiss, too, would have contained numerous microcline porphyroblasts. Further, I shall mention the existence of sparse eruptive dikes of Askim granite not only in rather resistant basites but also in more acid rocks (femic gneiss especially, see Fig. zo), and even close to the contacts the latter are free from microcline eyes. These data evidence that the early eyes derive from the same magma as the granitic matrix enclosing them. They are penetrative, and in part they even seem to be less influenced by the syn-orogenie stress than the matrix. Accordingly, the latter should be interpreted as the principal product of the crystallizing magma, whereas the eyes should have taken at least most of its potassium from the residual magmatic solutions. The remaining components of the eyes should derive from the original mineral individuals that have been replaced by the growing microcline. We can thus define the early microcline eyes as products of deuteric potassic metasomatism. Several remnants of plagioclase in the early eyes evidence the correctness of this statement. Plate r shows that the borders between Askim granite and non-porphyritic late Gothian granites are sometimes distinct, sometimes diffuse. The frequency of diffuse borders, viz. transitional granites, obviously to a high degree depends on the frequency of late eyes (compare the outcrop SE of Släp above described). It certainly also depends, however, on the mode of the magmatic evolution of g ranites in various parts of the Gothenburg-Onsala fold. Early crystallization of abasic granitic magma portion low in potassium, followed by late crystalliza­ tion of an intermediate granitic magma portion rather high in potassium within the same part of the fold, thus inclines to create distinct contacts be­ tween the resultant non-porphyritic (basic) and porphyritic (intermediate) gra­ nites. On the other hand, variable frequency of residual solutions rich in po­ tassium within one crystallizing mass of intermediate granite inclines to develop plagioclase-granite, transitional granites, and Askim granite. Such a frequency PER H . L UNDE GÅRDH.

Fig. 2r. Intrusion of late fine·grained granite in Frölunda granite with dike of basic granite (to the left) and remnant of amphiboli tic gneiss (upper right corner). S of Näset , parish of V. Frölunda. Photo by P. H . Lundegårdh 1952.

variation seerus to be due to a ca-operation of crystallization differentiation and squeeze developed by the syn-orogenie stress. A sample of typical non-deformed Askim granite from the Billdal-Kullavik area has been subjected to chemical analysis (Table II). As is reflected by the contents of sodium and potassium respectively, the matrix is rather sodic, viz. poor in microcline. On the whole, the rock might thus be characterized as a plagioclase-granite with microcline eyes. The latter are penetrative and fre­ quently enclose remnants of plagioclase. The mineral composition is as follows: plagioclase (with sericite) > microcline :::::: quartz > biotite ~ epidote (with clinozoisite) > titanite ~ apatite > magnetite (titaniferous) andfor titanomag­ netite ~ zircon :::::: penninite :::::: museavite > pseudo-allanite.

Tb e late K arelian granite S of Näset has already been mentioned now and then. This rock has intruded into Frölunda granite (Plate r ). Its borders are always distinct and diseardant (Fig. 2 1 ). If often contains small xenoliths of met a-basites, and dikes of final Karelian pegmatite (Fig. 22 and p. 49) are sometimes found. The general behaviour of the Näset rock is quite similar to that of the Bohus granite. Regional studies and comparisons have also shown tbat the granite S of Näset as well as the pegmatite cutting it belong to the distal products of the lat e Karelian migmatization. PETROLOGY OF THE ~IÖL N DAL -STYRSÖ-YALLDA REGION. 47

The granite S of Näset is grey, fine-grained and non-deformed. It has the following mineral composition: quartz > microcline > oligoclase > biotite ~ muscovite ::::=: epidote ~ titanite ~ apatite ~ pseudo-allanite. Titanite and apa­ tite are minor minerals, pseudo-allanite is an accidental constituent. The microcline is penetrative. The largest mineral individuals of the rock are cam­ posed of either microcline (sometimes perthitic) or quartz, though small grains of both are also frequent. The oligoclase has in part been sericitized. It often contains deuteric muscovite and epidote. The titanite has sometimes altered to a dense mass of minute grains. The considerable epidote content of the late Karelian granite indicates strong alteration effected by residua! solutions. These have probably also penetrated the surmunding Frölunda granite and may have there developed not only epidote (compare p. 39) but even eyes of microcline (see p. 44).

Pegmatite, Aplite, Veined Gneisses.

In the Mölndal-Styrsö-Vallda region, the coarse pegmatite and its inferior sugar-grained companion aplite as a rule appear as irregular masses, minor dikes, glands, veins, and schlieren. In the northern part of the region, however, their mother solutions have sometimes been able to produce larger dikes. The colour of the pegmatite passes from grey-white to various shades of red. Red-grey-white tints are most common. The aplite is generally pink or pale red-grey. The principal minerals of both rocks are grey quartz, grey-white, pink or red microcline, and white sodic plagioclase. In the pegmatite, some mica is also frequently present. Most microcline individuals examined are perthitic. In the region considered, two groups of pegmatite have been distinguished, viz. an older one of Gothian age and a younger one of late Karelian age. The Gothian pegmatite dominates. It forms both intrusions in and part of the veined gneisses (Plate r, p. II ff. and Fig. 5), whereas the Karelian pegmatite always appears as distinct dikes (Fig. 22). The Gothian pegmatite is con­ centrated in the eastern part of the Kållered-Mölndal area and in the Styrsö archipelago, while the Karelian pegmatite has only been observed in the V. Frölunda-Askim area and in the western part of the Mölndal area. As I have earlier touched upon (p. 6 and Table I), the Gothian pegmatite has been divided in to two generations. The older of these seems to be represented among the xenoliths of the plagioclase-porphyrite (p. 32) and should have developed during an earl y Gothian migmatization of the lowest ( = outermost) strata of the ancient Gothenburg-Onsala syncline. Owing to the extensive alterations effected by the final Gothian migmatization (see below), the areal distribution of the early pegmatite is unknown, however. The younger, and most important, generation of Gothian pegmatite is characterized by a frequent content of magnetite, which may at times grow considerable and which has most probably been inherited from dissolved acid gneiss (compare p. 25). Garnet , too, has been observed. This secondary peg- PER H . L UKDEGAR DI-I.

F ig. 22. Dike of latest pegmatite in late fine-grained granite. S of Käset, parish of V. Frölunda . Photo b y P. H . Lnndegårdh r 952. matite is the principal product of the final Gothian migmatization, which has been described in my Onsala paper (P. H . Lundegårdh rgsr, pp. r88- r gr). I have interpreted this process as due to an eastern-western strain that eaused glidings of blocks (the Mölndal-Kungsbacka fault, for instance) and opened the coneardant sheets of the outer and older supra-cm st al rocks of the Gothenburg-Onsala fold. These became thus migmatizable, viz. permeable to rising solutions of acid silicates (see p. II ff. and Fig. s). On the other hand, the massive granitic kernel of the fold with its huge bu t soldered xenoliths and strokes of younger supra-cm stal rocks has been less susceptible to altering actions. Accordingly, we have there only dikes of lat e Gothian pegmatite, no real veined gneisses. During the migmatization, part of t he outer strat a of the Gothenburg­ Onsala fold became plastical. Glidings frequently disjointed more rigid rocks (Fig. 6; see also P . H. Lundegårdh rgs r, Figs. II, rz, and 14). F urthermore, the acid strat a were in part dissolved and mobilized. They have originated the intmsive granite dikes exemplified by Fig. 23 and described on p. 40. In the gneisses of the inner parts of the Gothenburg-Onsala fold, minor pegmatitic and aplitic schlieren and veins are sometimes met with. These have developed contemporaneously with the late Gothian granites. Their composi- PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 49

Fig. 23. Di scorrlant granitic dike of dissolved and mobilized acid gneiss in grcy plagioclase­ granite with shcet of amphibolite (around the compass). Tjurholmen (islet WNW of Kullavik ), parish of Slä p. Photo b y P. H. Lundegårclh rgsr. tion clepends upon the character of the mother rocks. Conelitians being favour­ ahle, they can thus even contain hornblende. The late Karelian pegmatite is intimately related to the granite S of Näset (see ahove), which it has been seen to cut (Vig. 22). It often contains green microcline (amazonstone) and fluorite (green and v iolet). The content of cer­ tain rare minerals - beryl, columbite, and monazite, is sometimes consider­ able. The most important locality is the Högsbo intrusion - a pair of dikes situated 2 km NE of V. Frölunda church, just outside the northern border of Plate r. The Högsbo pegmatite has been investigated by N. Sundius (r gso, p. 473) . He reports that the rock has intruded as two paraHel dikes (width = 6 and r2- r5 m respectively), the strike and clip of which are N7o 0 W, 6o 0 SSW. (It should be noted that part of the dolerite dikes described in the next chapter show the same orientation.) The plagioclase and mica (muscovite preferentially) is concentrated in the marginal zones of the dikes, whereas the inner parts are essentially camposed of quartz and pale red, perthitic microcline. Crystals of the latter mineral measuring as much as 2 m in length have been observed. Veins of white albite especially, and quartz, sometimes occur in the perthitic microcline. In the kernel of the larger dike, violet and green fluorite seems to have been concentrated. Furthermore, beryl, columbite, and monazite have cryst allized in the inner part of this dike. The specific gravity of the Högsbo columhite and monazite amounts to 5·39 and 4.8 respectivelv.

4 ---.'i30834. S. C. U. , Ser. C. _-\r:t> 5_1r. L u nd~ ~ · å rd/i _ so PER H. LUNDEGÅ RDII.

In his Högsbo paper, Sundius also mentians that columbite has only been found in Sweden at Timmerhult on Orust, Central Bohuslän, and Varuträsk in Vesterbotten, Northern Sweden. Monazite is known from L. Holma in Lur, Northern Bohuslän, and Kårarvet at Falun, Dalecarlia. Of these four occur­ rences, the two in Bohuslän are intrusions of lat e Karelian pegmatite. Even with respect to its content of rare minerals, the Högsbo pegmatite should thus be classed as late Karelian. Moreover, the absence of magnetite and the presence of the rare minerals mentioned, evidence the distal, or rather, epi­ orogenic character of the Högsbo pegmatite.

Basic Dike Rocks. Early Diabase. Between Näset and Fiskeb äck, on In-Vinga N of Vin ga, and in the n orthem part of the Styrsö archipelago, especially on Rivö and Galtö (Plate r), swarms of dikes of green-black to black, fine-grained di abase cut the veined gneisses and late Gothian granite (Fig. 24). The dikes are quite narrm.v (Fig. 24). As a rule; their width varies from z cm to 3 m. The primary ophitic texture of the diabase has been preserved. The marginal zones of some dikes are, however, schistase and amphibolitic. The strike of most dikes is N- NE. The dip falls steeply towards W- NW. The diabase considered is essentially camposed of corroded primary laths of andesine and of various deuteric mafic minerals, green uralitic hornblende and biotite preferentially, further epidote and chlorite. Titanite is often a rather important constituent , whereas titaniferous magnetite, titanomagnetitf', apatite, and pyrite should be classified as minor minerals. A sample of early diabase from one of the dikes between Fiskebäck and Näset has been subjected to spectral analysis (Table III). The data obtain ed will be disenssed on p. s6. In the centre of Rivö, the kernel zone of one diabase dike contains small xenoliths of quartzite and altered slate gneiss belonging to the early Gothian supra-cmstal series. Most xenoliths display a mass of fine-grairred granoblastic, or coarse to medium-grairred quartz with blebs of plagioclase and crystals of epidote. In some xenoliths, accidental microcline has been observed. H . E. Johansson (r93 I , p. 45) considers the dikes now described to be equiv­ alent to the Koster diabase and dolerite. On the Koster isles in Northern Bohuslän, this group of rocks forms several hundreds of dikes, the strike of which curves from NNE (in S) to N and NNW (in N). B. Asklund, who has mapped and investigated the Koster rocks, mentians that part of the minor dikes and the margins of part of the major dikes have become schistase and amphibolized. The undisturbed and wider dikes display a dolerite camposed of labradorite and augite preferentially, further some hypersthene, and at times also olivine (Asklund rgso, p. 53). Asklund parallels the Koster di abase and dol erite with the hyperite of Central Southern Swecl en. As th e Koster dikes PETlWLOGY UF THE MÖLN UAL- STYHSÖ- VALLDA REGlON. 5I

Fig. 24. Dikes of early diabase in mixed gneiss in part transformed into Frölunda granite. Northern In-V inga, N of V inga, pari sh of Styrsö. P ho to b y P. H . Lundegårdh 1952. have been cut by the Bohus pegmatite (A. Gavelin rg14; B. Asklund rgso, pp. 53- 54) and the Rivö-Gal tö dikes are younger than the fin al Gothian migmatization, the resultant generation of basites should have developed in Karelian time.

Late Diabase and Dolerite. The youngest magma that intruded into the Mölndal-Styrsö-Vallda region has åccupicd the space offered by an opening system of steep western to west­ north-western and northern-southern joints. The resultant diabase and dolerite dikes have been observed in the southern part of the Styrsö archipelago and in the central and southern parts of the mainland (Plate r). North of the mapped region, a number of dikes cut the Gothenburg district and the parish of Torsby (see H. E. Johansson rg3r, Plate r). The western to west-north-western joints existed and even in part opened in latest Karelian time, as exemplified by the Högsbo pegmatite dikes (see above). The northern-southern joints developed during the late Gothian tectonization. The width of the diabase and dolerite dikes accessible for determinations varies from 2 dm to 25 m. The largest dikes in the region investigated pass 1 r / 2 km S of Billdal (the H eden dike) and through Kullavik (the K yvik dike). The maximum width of the former amounts to 25 m at the locality shown in Fig. 25, while the maximum width of the latter is ro m. Though rather short. 1 the perpendicular dike 5 / 2 km E of Billdal has a considerable width, too. 52 PEH I-I. LUNDEGÅHDH.

The morphological appearance of t he late diabase and dolerite dikes is characteristic. In certain places the rock has been quite resistant to weather­ ing and stanels up as hills or walls (Fig. 25), in othersit has weathered to beds of gravels or valleys covered with moraine etc. and thus cannot be detected at all. As is seen from Plate r, both the Heden and Kyvik dikes have been cut and dislocated by tectonic movements along a northern-southern zone of parallel joints. This zone was developed in late Gothian time, contemporaneously with the great Kungsbacka-Mölndal zone, with which i t also runs paralle l. Movements have here certainly occurred in late Kareli an time and probably in Permian time, too (campare the extensive Permian magmatic activity in the Oslo field NNW of t he Gothenburg-Onsala fold). The kernel zones of the wieler dikes display a fresh, black or grey-black do le­ rite, and the same holels for some of the narrow dikes. The northern-southern dike SW of Valida church is thus camposed of a well-preservecl dolerite which is even rather meclium-grained, though the dike is quite narrow (z m). ln normal cases, however, the thinner dikes are wholly fine-grain ecl , whereas the wieler dikes show fine-grained margins and fin e- to medium-grained kerncls. The dolerite, and the diabase, too, shows a beautiful ophitic clevelopment with laths of plagioclase which are as a rule margin ally corrodcd in the Jatter rock. \Veak partial sericitization is met with now and then. Zonal extinction has been observcd in many crystals. The composition of the plagioclase as a r ule varies from 40 to 6o % anorthite, though varieties with more acid plagio­ clase also exist. The main mafic minerals of the dolerite are pyroxene (often ophitic), further titaniferous magnetite and frequently oli vine, too. The dolerite of the widest part of the Heden dike (Fig. 25) thus shows the following mineral composition: 1 andesine (about 40 % anorthite ) > olivine (wit h serpentine) 2:: cl inopyroxene > titaniferous magnetite. Biotite (deuteric; see below) should be classed as an inferior constituent. Quartz (late) > calcite (deuteric) 2:: apatite are minor minerals. This rock has been subjected to chemical and spectral analysis 0 (Tables II- III and p. 56). Its olivine shows z Vy = go , or samewhat more, its 0 clinopyroxene (pigeonite) yl\c 28- 33° and zVy < 6o • Birefringence and in­ Jices of refraction increase on increasing values of yl\c. A sample from the Kyvik dike about z km E of Kullavi k is camposed of andesine-labraclorite (50 % anorthite) ::?: clinopyroxene with deuteric altera­ tion proclucts > titaniferous magnetite. Apati te and q uartz (late) are minor minerals. The clinopyroxene, mainly an augite (yl\c = 41 - 45 °), has to a !arge extent alterecl to a dense mass of biotite, uralite, chlorite etc. The perpendicular dike SW of Valida (see above) displays labradorite (55 - 60 % anorthite) > chlorite + serpentine > clinopyroxene > olivine ~ m ag­ netite. Biotite has here to be classified as a minor mineral, whereas apatite is rare. This is an early and unusually basic modification with no quartz, very little apatite, a rather consiclerable content of nickel (see Table III and

1 V/hen borderin rr upon quartz, t he plagioclasc frcquenll y grows rnorc acid. I'ETlWLOL;Y 0 1' THE MÖLNDAL- S'l' YHSÖ- VA LLJJA HEGlON. 53

1 Fig. 25 . Hcsistaut dike o f late doleritc. At H ällesås 4 / 2 km ESE of Billdal, parish of Liudutue. Photo b y P . H. L undcgårtlh rg sr. p. 56) , and an olivine (zVy = 85°) richer in magnesium than that of the Heden dolerite (zVy = 90°, or somewhat more). 1 On the isle of Stora Ryggholmen 2 / 2 km NW of Kullavik, the dolerite con­ tains both clino- and orthopyroxene (hypersthene). The main mafic minerals of the late diabase are uralite, chlorite (with ser­ pentine), and biotite. These have all developed by autometamorphism (in part autometasomatism, too). The uralite as a rule shows pale pleochroism in various shades of green. During the alteration of the original mafic minerals, the mar­ gins of the plagioclase laths have as a rule become corroded. In certain samples of diabase, small quantities of epidote, pyrite, and calcite have been observed. In my Onsala paper (P. H. Lundegårdh 1951, p. 194), I suggested that the late diabase and clolerite dikes of the Gothenburg-Onsala fold might be either pre-Hercynian or Lower Permian. The form er proposal is supportecl by the presence of wide beds of early post-Silurian plateau-clolerite in Vestergötlancl. B. Asklund (1950, p. 56) reports sparse finels of eastern-western to east-north­ eastern clolerite dikes in the Koster archipelago. These dikes are similar to those met with S of Gothenburg and thus even in part olivine-bearing. Asklund is inclined to paraHel his eastern-western dikes with the Vestgöta plateau­ clolerite, too. In Scania, several dolerite dikes running towarcls WNW have, however, been interpreted as Lower Permian by S. Hjelmqvist (1939) . This investigator is inclined to consicler the opening joints thus filled with rising basic magma as cleveloped by the Hercynian orogenesis. In his tectonic and morphological monograph on the Swedish Skagerack coast (Bohuslän especially) , E. Ljungner (1927, pp. II1- II5) gives a survey 54 P E R .H. I. UN UEGÅHDH.

of the west-north-wes tern diabase and dolerite dikes of South-Western Sweden. Ljungner here interprets themas Algonkian, and among petrölogical parallels he mentians the Jatnian Åsby dolerite of Central Sweden and the Billingsfors dike in Dalsland. Later in his monograph, Ljungner (1927, p. 249) seems in­ clined to consicler the WNW-dikes to be olderthan the Bohus granite, which they have never been seen to cut. In a recent paper (Ljungner 1953, p. II2}, he stresses this circumstance oncc again. He has, however, as weil as Ask­ lund (rgso), suggestecl an Algonkian agc of the Bohus granite, too (Ljungner IY27, p . 24Y)· Now it has to be pointcd out, t hat t he WNW-clikcs appcar as swarms. The Seanian dikes form such a swarm, the Gothenburgian ones another swarm, and those of the Kostel isles a thircl swarm. The Eillingsfors dike in Dalsland belongs to a fourth swarm (campare W. Larsson 1949). Between Scania and the Gothenburg-Onsala fale!, no WNW-dikes have been yet observed, in spite of W. Larsson's cl etailed and careful mapping of the southern part of this region (unpublisbed maps kindly placed at my clisposal) . And the rocks of Southern Halland are early Gothian, if not in part still older. Ljungner's objections against a Palaeozoic age of the WNW-clikes thus do not seem to be decisive, though, on considering the petrology and mode of occur· rence of the WNW-dikes of Gothenburg, Koster , and Dalsland, I do not feel inclined to parallel them with the Vestgöta dolerite. They are certainly younger than the Bohus granite, but I join Ljungner in classing them as Algonkian. The dislocations of the Heden and Kyvik dikes seem to have occurred in Permian time, simultaneausly with the magmatic activity in the Oslo fi eld.

Geochemical Evidences.

Typical representatives of the basic rocks of the Gothenburg-Onsala fold have been analysed spectrographically by means of the continuous are method. The elements determined are chromium, cobalt, and nickel. The accuracy of determinations of this kind has been estimated in P . H. Lundegårdh 1946, on pp. 22 - 23. As originally statecl by V. M. Golclschmiclt (see 1945, pp. 3- 4), chromium especially, and nickel have been enriched in magmatic first-differentiates. The validity of this rule has been evidenced by numerans gcochemical investigations (see P. H . Lundegårdh 1949). On studying the petrological consequences of this · rule, I have found (op. cit., p. 6) that cobalt is normally uniformly distributed in the various stages of basic magmatic differentiation (except the final pro­ clucts, where even this meta! vanishes tagether with magnesium). Accorcling to the data given in Table III, the Gothian effusive simatic derivatives (basic volcanics) are rather rich in chromium - the agglomeratic basaltic tuff near Meryt contains as much as 400 p. p. m. (parts per million) , whereas the infra-crustal simatic rocks contain far less chromium - a homo- PETROLOGY OF THE MÖLNDAL-STYRSÖ-VALLDA REGION. 55

Table III. Distribut-ion of chro1nimn, cobalt, and nickel in the basic rocks of the M ölndal-Styrsö-Onsala region, in order from youngest to oldest. Analysts: J. Raudsepp (]. R.), V. Muld (V. M.), and P. H . Lunclegårclh (P. H. L.).

1 J

~ ~:v~::~me ta-peridotit e~:~ l St~~~~~;~~nSi~r~::_:: of- :e l __:~_!_ __ ! ~ .l 270 l

1 See P. H. Lundegårdh Ig5I, Plate r. 2 May be oleler than the stensholmen clavainite (see helow). geneous ultra-basic member of the series (uralite-porphyrite WNW of Meryt) shows only rso p. p. m. Moreover, the basaltic tuff seems to be quite an un­ differentiated simatic clerivative,l whereas the uralite-porphyrite is an early product of basic magmatic clifferentiation and woulcl thus be expected to contain more chromium than the former. Indeed, the parental simatic magma of the infra-crustal basites seems to have differentiated to some extent before the development of the bodies now visible, in the same manner as the magma of the early generation of ultra-basic gabbro in Central Roslagen NE of Stock­ holm (P. H. Lundegårdh 1949, p. zo). On camparing the clavainitic bands of the stensholmen gabbro NW of Säre with the uralite-porphyrite WNW of Meryt, we find that the former are prod­ ucts of rhythmic crystallization differentiation of a magma. They have been originally camposed essentially of olivine enriched in nickel. In contrast to chromium, this metal easily enters the olivine lattice. The plagioclase-porphyrite of Vinga mineralogically displays a late-magmatic charader (p. 30), as is also evidenced by the chemical and spectrographical data - Mg: Fe = o.' 8, low contents of chromium and nickel. As I have stated

' S. Landergren's (rg5 1, Table r) geochemical investigations on Got hi an supra-crustal meta­ basites from the vicinity of Varberg in Ce ntral H alland h ave given quite similar results: Cr = 200- JOO, Co = 30- 40, Ni = go- TIO p.p.m. (Two samples of amphibolite.) PER H. LUNDEGÅRDI-I.

earlier (P. H. Lundegårdh 1950, pp. 5I- 52) , the quotient Mg:F e is > I in early products of basic magmatic differentiation, o. s- r in m idel ie products, and < o. 5 in late products. The biotite-norite NE of V. F rölunda, should also be classifi ed as a late basic differentiation product (compare p. 35). Here we find the relation Co = Ni > Cr. According to Goldschmidt's rules, the relation Co > Ni > Cr should hold for late products of differentiating basic magmas (P. H . Lundegårdh 1949, p. zr). The chromium content (65 p. p. m.) is rather high, however, as compared with other Swedish late hasic roc ks of magmatic origin (see P . H. Lunclr­ gårdh 1949). The early diabase that cuts the bed-rock between Fiskebäck and Näset cliffers from the Koster cli abase in having much more chromium (65 p. p. m.) and nickel (z 6o p. p. m.). Two general samples of the latter show Cr = s - ro, Co = 45 - 50, and Ni = 85-90 p. p. m. (P. H. Lundegårdh 1949, Table {_), p. 4fi .) Two samples of olivine-bearing late clolerite have also been analysed. Thesf' show the same trend as the n earest Palaeozoic plateau-dolerite of Vestergötlancl , 11 i z. that of Hunneberg (Mount Hunne). Two representat ivesamples of central and margin al Hunne dolerite contain 40- fio p. p. m. of Cr, 40-50 p. p. m. of Co, and 70 p. p. m. of Ni (analyst: J. Raudsepp). This is, however, the normal trend of olivine-bearing basites that are not first-differentiates (see P. H. Lundegårdh 1949) . Besides, an analysis of a general sample of the more distant Kinne dolerite displays quite different values: Cr = 6, Co = 400, and Ni = zoo p . p. m. (general sample; P. H . Lundegårdh 1949, Table g, p. 46). The trend encountered in the late clolerite S of Goth m bnrg cannot thus he consicl err rl ;;s incli cativP of a Palaeozoic a.ge of t hi s rock .

Summary of Genetical Conceptions.

During the late Gothian (Table I ) deep-folding and heating of the synclinal Gothenburg-Onsala volcanics and sediments, pressure and space conelitians were quite unfavourable to the intrusion of large juvenile magma portions (compare H . G. Backlund 1936). The rather sparse finels of plutonic gabbroic­ dioritic rocks evicl ence the correctness of this statement. The mineralizers given off by the cryst a1lizing mother magma of these basic rocks and the underlying sim a were dissolved in the water of the sediments and immediately attacked the supra-crustal complex. The deeper the complex was fo lded, the greater became the heat supply, and thus not only granitizations in situ were effected (Frölunda granite, microcline-granite) bu t even dissolutions and mobilizations of supra-crustal rocks (secondary magmas later congealing as basic, plagioclasic, and Askim granites). The mineral components of the ancient supra-crustal rocks encl osing the clisjointecl beds of meta-basites within, for example, thr K1Illavik-Rilld al area are thns at prrsent components of srconcla ry granitrs. I'ETROLOGY OF THE MÖLN DAL--'STYRSÖ-VALLDA REGION.

In the outer parts of the Gothenburg-Onsala fold, the supra-crustal layers remairred coneardant and mostly escaped granitization. During the final Gothian migmatization, these schistase coneardant layers opened, however, owing to lat eral strain, and thus mineralizers could rise once again and start their destructive work. Veined gneisses and intrusions of secondary pegmatite are the visible results. Part> of the secondary mother solution of the pegmatite has developed on dissolution of alkaline gneiss rich in magnetite. Accordingly, the pegmatite bears ri;Iagnetite. In late Karelian time, wide parts of the bed-rock of South-Western Sweden were subj ected to a new orogenesis. Stress and strain worked. Folding, faulting, and migmatizations occutred. The palingenic rocks intruding during the final phase of this orogenesis seem to have an origin both distant and dif­ ferent from that of the late Gotnian migmatization products, however, since the resulting pegmatite contains minerals that are lacking in the late Gothian pegmatite. This is also in agreep1ent with W. Larsson's (1947, Fig. r) concep­ tion of a late Karelian zone o( palingenesis that has been deep-seat ed even in relation to the pre;=;e,nt surfa::e of the crust. Both in Karelian and post-Kareliarr time, systems of joints opened and allowed basic magma to rise and congeal. In the Gothenburg-Onsala fold, the visible results are two generations of diabase and dolerite, the younger of which is supposed to be Algonkian.

S-5:10884, S. G. U., So·. C. N:o .f.JI . Ltmdeg!l ,d!t,